Skip to main content

Full text of "Genesis and Development of Tantra"

See other formats

Genesis and Development 
of Tantrism 

Edited by 

Shingo ElNOO 

Institute of Oriental Culture 
University of Tokyo 

Genesis and Development 
of Tantrism 

Institute of Oriental Culture Special Series, 23 

Shingo Einoo is Professor of Indology at the 
Institute of Oriental Culture, University of Tokyo 


All rights reserved 

The authors: 

Shingo ElNOO 
Dominic GOODALL 

Taiken Kyuma 
Alexis Sanderson 
Francesco Sferra 
Tsunehiko SUGIKI 

Kimiaki TANAKA 

University of Tokyo 

Pondicherry Centre of the Ecole francaise 


Mie University 

All Souls College, Oxford 

Universita degli Studi di Napoli "LOrientale" 

Waseda Institute for Advanced Study, Waseda 


The Eastern Institute 

University of Tokyo 

ISBN 978-4-903235-08-0 
Printed in Japan 


Preface 9 

Introduction 11 

Shingo ElNOO : From kamas to siddhis — Tendencies in the Development 

of Ritual towards Tantrism — 17 

Alexis SANDERSON : The Saiva Age — The Rise and Dominance of Saivism 

during the Early Medieval Period — 41 

The Dominance of Saivism 44 

The Incorporation of Saktism 45 

The Etiolation and Subsumption of the Cult of the Sun-God . 53 
The Decline ofVaisnavism and the Rise of the Tantric 

Pancaratra Following Saiva Models 58 

Royal Patronage of Buddhism 70 

The Visnukundis of Andhra 70 

The Maitrakas ofValabhl 72 

The Karkotas of Kashmir 73 

The Licchavis of Nepal 74 

The Thakurl Kings of Nepal 77 

The Bhauma-Karas of Orissa 80 

The Candras of South-East Bengal 83 

The Khadgas of Samatata 83 

The Candras of Arakan and Miscellaneous Other 

Buddhist Kings of Eastern India 84 

The Pala Emperors and the Great Monasteries 

of Eastern India 87 

The Palas' Engagement with Saivism 108 

Buddhist Kings of Eastern India and their Commitment 

to Brahmanism 115 

Joint Patronage of Buddhism and Saivism in the Kingdoms 

of the Khmers, Chams, and Javanese 117 

The Development of Tantric Buddhism Through the Adoption 

and Adaptation of Saiva and Sakta Saiva Models 124 

The Parallel Repertoire of Rituals 124 

The Mahavairocanabhisambodhi, the Mahjusriya- 
mulakalpa, and Buddhaguhya 128 


Genesis and Development of Tantrism 

The Sarvatathagatattvasamgraha and the First Inroads 
ofSakta Saivism: Possession, Goddesses, 

and the Sacralization of Sex 132 

The Guhyasamaja: Copulating Deities, Sexual Initiation 

Rites, and the Sacralization of Impurity 141 

The Sarvabuddhasamayogadakinijalasamvara: Heruka 
and his Yoginis, Kapalika iconography, the Ganamandalam, 
and the Beginning of Saiva-Buddhist Intertextuality . . . 145 
The Yoginltantras and the Full Appropriation 

of Vidyapltha Saivism 156 

Chronology and Provenance 158 

Samvara /Vajrarudra and Vajravarahl: 

The Transformation of Bhairava and his consort . . 169 

The Rise of the Goddess to Independence 173 

The Adoption of the Vidyapltha's Carya and Yoga . 179 
The Incorporation of Text-passages 

from the Vidyapltha 186 

Converting the Outsiders 220 

The Reflux of Buddhist Saktism into 

the Saktism of Bengal 240 

The Jains' Adaptation of the Saiva Mantrasastra 243 

Saivism in the Brahmanical Substrate 249 

The Causes of the Dominance of Saivism 252 

The Early Medieval Process 252 

Saivism and Monarchy 254 

Saivism and the Royal Temple 274 

Saivism and New Settlements 280 

Saivism and Irrigation 282 

Saivism and Social Integration 284 

The Saiva-brahmanical Order 301 

Abbreviations 304 

References 305 

Conventions in the Footnotes 348 

Dominic GOODALL : Who is Candesa ? 351 

Kimiaki TANAKA : Nagabodhi's Srl-guhyasamajamandalopayika-vimsati-vidhi 

— The Sanskrit Text Restored from the Vajracaryanayottama — 425 

Francesco SFERRA : The Laud of the Chosen Deity, the First Chapter of the 

Hevajratantrapindarthatlka by Vajragarbha 435 

Taiken KYUMA : Superiority of Vajrayana — Part I: Some Remarks on the 


Vajrayanantadvayanirakarana (rDo rje thegpa'i mtha' gnis sel ba) 
Ascribed to Jnanasri — 469 

Ryugen TANEMURA : Superiority of Vajraydna — Part II: Superiority of the 
Tantric Practice Taught in the *Vajrayanantadvayanirakarana (rDo rje 
theg pa'i mtha' gnis sel ba) — 487 

Tsunehiko SUGIKI : The Structure and Traditions of the Systems of Holy Sites 
in Buddhist Samvara Cycle and Its Related Scriptural Cycles in Early 
Medieval South Asia — The Geography of Esoteric Buddhism in the Eyes of 
the Compilers of the Scriptures — 515 



The present volume is the outcome of an international workshop on 
Tantrism held on October 3-4, 2005 at the Institute of Oriental Culture, the 
University of Tokyo. The speakers and their titles were as follows: 

Shingo Einoo: 

Yuko Yokochi: 

Dominic Goodall: 
Francesco Sferra: 

Alexis Sanderson: 

Taiken Kyuma: 

Ryugen Tanemura: 

Harunaga Isaacson: 
Tsunehiko Sugiki: 

Kimiaki Tanaka: 

From kdmas to siddhis: Tendencies in the Development of 
Ritual towards Tantrism 

The Local Goddess Worship in an Early Saiva Centre 
Kotivarsa, Devikota or Sonitapura 
Who is Candesvara? 

Constructing the Wheel of Time: Strategies for Establish- 
ing a Tradition 

The Saiva Age: An Explanation of the Rise and Dominance 
of Saivism during the Early Medieval Period 
Some Remarks on "rDo rje theg pa'i mtha' gnis sel ba (*Va- 
jrayanantadvayapoha)" Ascribed to Jnanasri 
Superiority of Tantric Performance or Post-initiatory Ob- 
servance (caryd) Taught in the *Vajrayanantadvayapoha 
of Jnanasri 

The System of Hevajra Practice Associated with Dombi- 

Theories on the Cycle of Time, Calendar, and For- 
tunetelling Introduced in Tantric Buddhism in the Classic 
Indian Context 

Nagabodhi's Sriguhyasamajamandalopayika-vimsati- 

vidhi: The Sanskrit Text Restored from the Vajracarya- 

Eight of the above-listed participants have contributed to the Proceedings. 

I must apologize to the contributors that it has taken more than three years 
to publish the proceedings of the workshop. I would like to express my gratitude 
to Dr. Ryugen Tanemura who has kindly worked to prepare the final version of 
this volume by TeX compiler. 


24 December 2008 

Shingo ElNOO 



Shingo ElNOO 

The Nepal-German Manuscript Preservation Project has recently made 
available many new manuscripts for the use of Indologists, and this has con- 
tributed to significant progress in Tantric studies. The articles introduced below, 
with the exception to that of Einoo, have to a greater or lesser extent made use 
of these materials. In this sense, this volume can claim to represent the newest 
research on the subject. 

The first article in the collection is entitled "From kdmas to siddhis: tenden- 
cies in the development of ritual towards Tantrism" (Shingo Einoo) Vedic rituals 
are generally performed to obtain certain desires, and were, for the people of 
ancient India, a reliable way to fulfill their objectives. Some Tantric rites too 
claim to bring about the attainment of wishes. The objectives of such rites are 
usually referred to as siddhis or supernatural powers and belong rather to the 
category of supernatural phenomena. Thus they seem to be considerably differ- 
ent from the types of wish people expected to gain from the Vedic rituals that 
still remained within the sphere of everyday life. In his article Einoo attempts 
to trace changes in the selection of objectives in Vedic rituals, and in the selec- 
tion of siddhis in Tantric rites by examining some Vedic and post-Vedic texts. 
First, he analyses the Rgvidhana in order to discover what purposes and effects 
were expected and then the Samavidhanabrahmana. After that he compares the 
results obtained from the analysis of these two vidhana texts with the kdmya 
rites given in the Gobhilagrhyasutra. Among the Tantric ritual texts he has 
selected the Vinasikhatantra as an example of a Hindu Tantric text and the 
Amoghapasakalparaja as a text of Buddhist Tantrism to take an overview of the 
tendency towards a development from obtaining desires (kdmas) in the Vedic 
texts to getting supernatural powers (sidhis) in the Tantric texts. 

The second article is "The Saiva age: an explanation of the rise and domi- 
nance of Saivism during the early medieval period" (Alexis Sanderson). To de- 
scribe it, I would like to draw upon the author's excellent summary provided 
for the workshop, which I reproduce here with slight changes. "The early me- 
dieval period, from the sixth to the thirteenth centuries, saw a decline in the 
role played by the Vaidikas in general and the Atharvavedins in particular in 
the religious ceremonies sponsored by the court. Kings continued to make land- 
grants to Vaidika brahmins in order to promote agricultural expansion and the 
cultural penetration of new territory, and they continued to impose and uphold 
the brahmanical social order, but their personal devotion shifted to the deities of 


Genesis and Development of Tantrism 

the initiatory religions that integrated the brahmanical tradition but claimed to 
rise beyond it, or to Mahayana Buddhism, especially in its Tantric development. 
Among these alternatives Saivism was the most widely favoured. In the decla- 
rations of religious adherence included with the titles attached in inscriptions to 
the names of rulers the epithet paramamahesvarah 'supremely devoted to Siva' 
is much the most frequent in this period, and of the many surviving temples 
established by rulers throughout the subcontinent and Southeast Asia from the 
late sixth century onwards those dedicated to the worship of Siva are much the 
most numerous. The dominance of Saivism is also manifest in the fact that the 
other main bidders for royal patronage, Buddhism, Pancaratrika Vaisnavism, 
and Jainism, as well as the earlier forms of Saivism itself, were fundamentally 
revised or expanded along the lines of the Saiva Mantramarga as they sought 
to maintain their hold on the sources of patronage. As for the other two cults 
that held the allegiance of kings during this period, those of the Goddess and the 
Sun-God, the former was progressively subsumed within Saivism, and the latter, 
though once equipped with its own canon of scriptures, suffered a similar fate." 
In his article Sanderson argues from ample textual and epigraphic evidence that 
Saivism rose to its position of dominance by expanding and adapting its reper- 
toire to contain a body of rituals and normative prescriptions that legitimated, 
empowered, or promoted the key elements of the social, political and economic 
process that in its various regional adaptations characterized the working of the 
state in the early medieval period. 

Dominic Goodall presents the third article "Who is Candesa?" Candesa is re- 
ferred to variously as Candesvara, Candikesvara, Candisa, Canda and, in Tamil 
sources, as Canti and Tanti and he is treated as a guardian to Saiva shrines, 
as a warrior leader of ganas, as the consumer of offerings that have been made 
to Siva, as the punisher of the transgressions of Saiva initiates, as Siva's agent 
in property transactions, as the transmitter of Saiva knowledge and as a super- 
bhakta who severed his own father's legs because of his father's impiety. Some 
evidence suggests that Candesvara is a form of Siva or a manifestation of his 
anger, but other evidence presents him as agana, as Siva's chief devotee or as his 
principal servant. Goodall admits that this figure emerges as a rather jumbled 
picture. He minutely discusses these various identities and concludes that Canda 
was at one time and for certain groups a form of Siva himself and not originally 
Saiddhantika. It is true that Candesa occurs only rarely in non-Saiddhantika 
tantric material. Candesa was, however, represented elsewhere and such repre- 
sentations have either been destroyed or have not been recognized because their 
iconographies were so different. The many undatable but early free-standing 
Candesas that are found in South India are independent of the Saiva Siddhanta. 



Candesa has thus been incorporated into the Siddhanta from an existing tradi- 
tion. His association with the Siddhanta today might well be the result of the 
Siddhanta's attempt to gradually appropriate to itself the role of decreeing how 
temple worship should be conducted in South India and the nirmalya-bearing 
Candesa in turn was from the first a Saiddhantika figure, because food offerings 
to the linga became invested with terrible power and had therefore to be con- 
sumed by an especially fierce form of Siva himself. His flourishing in the South 
was in part a consequence of the popularity of the legend known to Tamil sources 
as the transformation of the shepherd Vicarasarman into the mr/na^ya-receiving 
Candesa, the first servant of the Lord and the archetype of the Mahesvara devo- 
tee. Goodall's discussion is corroborated not only by textual evidence but also 
forty -four photographs of Candesa taken from various parts of India. These pic- 
tures, mostly in colour, are collected in the beginning of this volume. 

The fourth article "Nagabodhi's Sri-guhyasamaja- mandalopayika-vimsati- 
vidhi: The Sanskrit text restored from the Vajracaryanayottama" (Kimiaki 
Tanaka) is concerned with the Sri-guhyasamajamandalopayika-vimsati-vidhi, 
attributed to Nagabodhi, who belonged to the Arya school of interpretation of the 
Guhyasamajatantra. It is a ritual manual for the thirty-two-deity mandala of 
the Guhyasamajatantra with Aksobhyavajra in its centre and is one of the basic 
texts for mandala rites in late Tantric Buddhism. This important text has been 
long missing but Tanaka has found that the Vajracaryanayottama includes a 
work combining this text with explanatory comments. He has already published 
all the chapters of the Vimsatividhi recovered from the Vajracaryanayottama 
in various other publications, and is now preparing to combine these separate 
chapters into a monograph with the present paper serving as a general intro- 
duction to the text. He states that this text can be dated to the period from the 
middle of the 8 th century to the early 9 th century, when Tantric Buddhism was 
introduced to Tibet for the first time. 

The fifth article "The laud of the chosen deity, the first chapter of the 
Hevajratantrapindarthatika by Vajragarbha" (Francesco Sferra) is a new edition 
and English translation of the text one of the first works of the Kalacakra 
tradition. This chapter describes the characteristics of the true interpreter of 
the scriptures, establishes the hermeneutical criteria for their interpretation 
and gives an important role to the Adibuddha, which is believed to be the inula 
text of the Laghukalacakratantra. This chapter is therefore fundamental for 
an understanding of the founding strategies of the Kalacakra system and of its 
early masters. 

The sixth article is entitled "Superiority of Vajrayana, Part I: some remarks 
on the *Vajrayanantadvayanirakarana (rDo rje theg pa'i mtha' gnis sel ba) as- 


Genesis and Development of Tantrism 

cribed to Jnanasri" (Taiken Kyuma). In late Indian Buddhism one and the same 
author sometimes wrote both Tantric and non-Tantric works. In such cases it is 
not always clear how the author estimated Tantric and non-Tantric Buddhism. 
The text discussed by Kyuma offers a good example to clarify this problem, be- 
cause it is ascribed to Jnanasri, an author of works both on Tantric and non- 
Tantric Buddhism, and it teaches the superiority of Tantric Buddhism over non- 
Tantric Buddhism. In order to show how the author proves the superiority of 
Tantric Buddhism Kyuma gives a rather detailed synopsis of the text, largely de- 
voted to an explanation of the eleven means characteristic of Tantric Buddhism 
alone. Then he discusses in detail the identification of Jnanasri, because both 
Jnanasrimitra and Jnanasribhadra are known simply as Jnanasri; he concludes 
that the author of this text is certainly Jnanasrimitra. 

The *Vajrayanantadvayanirakarana demonstrated the superiority of 
Tantric Buddhism over non-Tantric Buddhism through the eleven kinds of 
skillful means, of which the eleventh insists that the Tantric practice is supe- 
rior because it rejects the three wrong practices: (1) practices which weaken 
the faculties of the practitioner, (2) practices which distract the mind of the 
practitioner, and (3) practices which produce the cognition that things are real 
and exist. In this way Jnanasri justifies the following three points: (1) Tantric 
Buddhism takes a negative attitude toward traditional Buddhist asceticism, (2) 
it distances itself from adherence to external religious acts such as the worship 
of a stupa and the recitation of a scripture, and (3) Tantric practice involves 
consumption of impure substances. These assertions can be found also in earlier 
scriptures and other scholastic treatises. By examining these statements and 
comparing them with the opinions of Jnanasri, Ryugen Tanemura attempts in 
his article, the seventh, "Superiority of Vajrayana, Part II: superiority of the 
tantric practice taught in the *Vajrayanantadvayanirakarana (rDo rje theg pa'i 
mtha' ghis sel bet)", to ascertain the position of Jnanasri among discussions of 
Tantric practices. 

The last article, "The structure and traditions of the systems of holy sites 
in Buddhist Samvara cycle and its related scriptural cycles in early medieval 
South Asia: the geography of esoteric Buddhism in the eyes of the compilers of 
the scriptures" (Tsunehiko Sugiki), discusses the post-Gupta era Buddhist de- 
velopment of an orthodox system of eight great sites closely related with eight 
great deeds performed by the Buddha, and the later introduction by esoteric 
Buddhism of new systems of holy sites seemingly from outside. The Samvara cy- 
cle, on which his article focuses, contains many scriptures and scholastic works 
and each of them describes the system of holy sites from various perspectives. 
Sugiki analyses the complicated materials by first classifying the descriptions of 



holy sites into four typological traditions: (1) twenty-four holy sites systematized 
on the basis of ten categories introduced from the Saiva text Tantrasadbhava 
into the Cakrasamvaratantra, (2) twenty-four holy sites systematized on the ba- 
sis of twelve categories introduced into the Hevajratantra and adopted into the 
Samvara cycle, partly related with the Kaula and Sakta tradition, (3) twenty- 
four holy sites without systematization introduced from the Kubjikamatatantra 
into the Vajradakatantra, and (4) seventy-two holy sites as residing places of 
seventy-two magical female beings introduced into the Dakarnavatantra and 
found also in the Kalacakratantra. According to his analysis these systems of 
holy sites have two levels and the first level, i.e. the system of practice, has 
three dimensions, namely (1) holy sites as geographical locations, (2) holy sites 
as a mandala to be drawn or to be visualized, and (3) holy sites as an internal 
mandala identical with the structure of the practitioner's body. On the other 
hand the second level consists of the aetiological myth of the twenty-four holy 
sites. Sugiki clarifies how the Samvara cycle in its various texts developed the 
systems of holy sites which had been introduced from outside into its particular 
esoteric Buddhist version by the involvement of the orthodox Mahayana Buddist 
doctrines and the internalization of external practices. 

As Sanderson's contribution clearly shows, many of the articles in this 
volume take into consideration both Hindu Tantrism and Buddhist Tantrism. 
Tantric studies have taken a new turn, where relationships and parallelism 
between different trends of Tantrism are examined extensively. The relationship 
between Hindu and Buddhist Tantrism will be the subject of a special panel 
organized by Goodall and Einoo at the 14 th World Sanskrit Conference to be held 
in Kyoto September 1-5, 2009. It is hoped that the new tendencies brought into 
the Tantric studies in this volume will bear even richer fruit at this conference. 


From kamas to siddhis 
— Tendencies in the Development of Ritual towards Tantrism — 

Shingo ElNOO 

1 Introductory remarks 

Vedic rituals are generally performed to obtain certain desires, of which the most 
common are the prosperity of progeny (prajd) and cattle ipasu). 1 There is a cer- 
tain group of Vedic rituals which are referred to as "kdmya". Those which are per- 
formed following the basic pattern of the new and full-moon sacrifice are called 
kdmyesti 2 and those performed according to the basic pattern of animal sacrifice 
are called kdmyapasu. 3 According to the analysis of W. Caland, the objectives 
expected of the kamyestis are: progeny, cattle, prosperity (pusti), dignity of the 
Brahmin (brahmavarcasa), gold, the position of royal chaplain (purodhd), well- 
being (bhuti), village, to conquer rivals (bhratrvya), sorcery (abhicdra), concord, 
for one who is cursed, for a dethroned king, to win a battle, longevity, against pos- 
session by demons, eyesight (caksus), against consumption (rdjayaksma), rain, 
expiation of ritual defects (prdyascitta), etc. (Caland 1908: VI-VII). 

There is a Vedic sacrifice called cdturmdsya, a ritual complex that consists of 
the vaisvadeva performed in spring, the varunapraghdsa performed in the rainy 
season and the sdkamedha performed in autumn (Einoo 1988). In a previous 
article I explored what purpose each constituent rite performed at the different 
seasons aimed at and came to the conclusion that "The vaisvadeva, the first rite 
of the cdturmdsya, assures that progeny and cattle are born safely and constantly 
and that they will grow well by means of ample food. But the life of human be- 
ings is never without danger; the god Varuna punishes the transgressions one 
commits, one must overcome conflict with rivals (bhratrvya) and Rudra sends 
damage to human beings and cattle, frequently without any cause. So the sacri- 
ficer appeases Varuna by performing the varunapraghdsa, overcomes his rivals 

1 In the Brahmanas the expression prajayd pasubhih prajdyate "progeny and cattle 
are born constantly" and other similar expressions occur very often. For a collection 
of such expressions, see Oertel 1994 (1943): 1552-1565. 

2 For the kamyestis, see Caland 1908 (1968). 

3 No detailed study exists for the kdmyapasus like that of the kamyestis. Du- 
mont 1962: 246-263 and Dumont 1969: 34-66 translate TB 3.6 and TB 2.8 which 
deal with the kdmyapasus. Srautakosa, Vol. 1, Sanskrit Section, Poona: Vaidika 
Samsodhana Mandala, 1958, pp. 606-697 and Srautakosa, Vol. 1, English Section, 
Part II, Poona: Vaidika Samsodhana Mandala, 1962, pp. 877-893 are the Sanskrit 
texts and their English translations respectively. 


Genesis and Development of Tantrism 

by means of the sdkamedha and calms the anger of Rudra by the tryambaka rite, 
which forms part of the sdkamedha. For human beings death is unavoidable but 
they seek to escape an untimely death. The mahdpitryajha, which also belongs 
to the sdkamedha, helps one live for one's full life span. Thus, the people of an- 
cient India expected from the Vedic rituals that they might live as peacefully as 
possible within nature, which could be both mild and unruly' (Einoo 1986: 1066) 
I assumed that the Vedic rituals were a reliable way for the people of ancient 
India to fulfill their objectives, so eagerly desired. But Tantric rites too claim to 
bring about the attainment of wishes. J. Torzbk, for example, analyzes the rites 
described in the Siddhayogesvarimata and classifies the objectives of the rites as 
siddhis in the following way: 

1. sdttvika siddhis: well-being / being well-fed (pusti I dpydyana); expiation / 
pacification (sdnti); [saving things] in case some disaster occurs (upasarge samut- 
panne); conquering death (mrtyumjaya); eloquence / poetic talent (kavitva); the 
ability to be infinitely small, big etc. (animddigundh); final release (moksa). 

2. rdjasa siddhis: subjugating people to one's will (vasyd); attracting people 
(esp. women, dkarsana); going to the underworld (pdtdlecaratvam); flying (khe- 
caratvam); disappearing (antardhdnam); "pill-siddhi" (a pill, put in the mouth, 
is said to make one invisible, gulikdsiddhi); and a siddhi with a magic wand and 
a bowl (siddhakdsthakamandalau). 

3. tdmasa siddhis, twelve kinds of black magic (abhicdras) listed in chapter 
24: murder (mdrana), expelling someone (uccdtana), annihilation (jambhana), 
paralysing (stambhana), benumbing (mohana), "nailing down" (kllana), taking 
away someone's speech (vdcdpahdra), making someone dumb (mukatva), deaf 
(bddhirya), blind (andhana), impotent (sandhlkarana), and changing one's form 
(rupasya parivartanam) (Torzsok 2000: 138-139). 

In the Siddhayogesvarimata, the objectives of the rites which J. Torzbk lists 
as siddhis (or supernatural powers), especially the rdjasa and the tdmasa sid- 
dhis, belong to the category of supernatural phenomena and seem to be consid- 
erably different from the types of wish people expected to gain from the Vedic 
rituals that still remained within the sphere of everyday life. 

In this article I attempt to trace changes in the selection of objectives in Vedic 
rituals, and in the selection of the siddhis in Tantric rites by examining some 
Vedic and post-Vedic texts. First, I analyse the Rgvidhana, which prescribes a 
great number of rites performed by magically using the hymns and verses of the 
Rgveda, in order to discover what purposes and effects were expected of them. 
I then examine the Samavidhanabrahmana, which likewise enjoins the magical 
use of the sdmans. After that I compare the results obtained from the analysis 
of these two vidhdna texts with a short description of kdmya rites given in the 


From kamas to siddhis 

Gobhilagrhyasutra, a Vedic domestic ritual text. Among the Tantric ritual texts I 
have selected the Vinasikhatantra as an example of a Hindu Tantric text and the 
Amoghapasakalparaja as a text of Buddhist Tantrism, and make an overview of 
the tendency towards a development from obtaining desires (kamas) in the Vedic 
texts to getting supernatural powers (siddhis) in the Tantric texts. 

2 Analysis of the rites of the Rgvidhana 

The Rgvidhana consists of four chapters. Rgvidh 1.1-78 4 forms the introductory 
part; descriptions of rites using RV 1.1 and so on begin from Rgvidh 1.79. Most of 
the rites are simple japas or recitations of certain hymns or verses of the RV, and 
homas or offerings of butter in fire. From time to time we come across pujas or 
the worship of deities peculiar to the post-Vedic ritual texts (Bhat 1987: 87-94). 
The prescriptions of the rites are usually mingled with statements about their 
effects and purposes: 

tan japan prayato nityam istan kamant samasnute /1.84coV 

medhakamo japen nityam juhuyac cajyam etaya /1.85cd/ 

"One who, being pure, always mutters these verses obtains the desires he wishes." 

"One who desires intellect should murmur it or should offer butter with it." 

Rgvidh 1.84cd and 1.85cd are statements concerning the effect of using RV 1.2— 

3 and RV 1.18.6 respectively. In most cases one effect is assigned to one ritual 
act, but sometimes one performance claims more than one result as the following 
example shows: 

sauparnani pavitrani siiktany ekadasabhyaset I vanchan putran pasun vittam su- 
vargam ayur anandhatam /1.106/ "One who wishes sons, cattle, wealth, heaven, 
longevity, and not being blind should repeat the eleven purifying sauparna hymns 
(RVKh 1.2-12)." 

In this case we can consider that six kinds of effects are mentioned. In this way 
we are able to collect 413 statements about the effects of the rites from 652 verses 
in four chapters of the Rgvidhana and classify them as follows: 

1-1-0 to live a full life span (ayusya): 34 5 

4 In this article I use Bhat, M.S., 1987, Vedic Tantrism: A Study of Rgvidhana of 
Saunaka with Text and Translation, Delhi: Motilal Banarsidass as the text of the 
Rgvidhana. In Bhat's text, the verses of each chapter are numbered differently 
from the numbers of verses given in the text of Rudolf Meyer, 1877, Rgvidhanam, 
Berolini: Typis A. W. Schadi. However, as Bhat also gives the numbers of sections 
which almost correspond to those of Meyer's edition, it is not difficult to find the old 
numberings of Meyer's edition. 

5 The number given after : in each item refers to the number of occurrences of the 


Genesis and Development of Tantrism 

To obtain a full life span (ayus): 11, to obtain long life (dlrgha ayus): 4, to 
live long (jyog jwet): 1, to live for a hundred years: 3, no fear from the god of 
death: 2 6 , not to become sick: 13 

1-1-1 to obtain eyesight (caksus): 3 

1-1-2 to conquer death: 6 7 

1-2-0 to remedy disease (bhaisajya): 9 

1-2-1 to remedy consumption (yaksman): 3 

1-2-2 improvement of digestion: 3 

1-2-3 to remedy poison: 3 

1-2-4 to obtain a medicinal herb: 1 

2-0 prosperity (paustika) or for various desires (kamya): 112 

to obtain desires (kama): 20, to obtain wealth (sri): 10, to obtain well-being, 
etc. (bhuti, rddhi, sukha): 6, to obtain one thousand (sahasra): 2, to obtain 
possessions (vitta): 4, to obtain riches (dhana): 16, to steal riches by killing an 
enemy: 2, to obtain other wealth (dravina, ratna): 2, to obtain gold: 4, to ob- 
tain a house (grha), a dwelling place (nivesa): 2, to obtain clothes: 3, to obtain 
food: 8, to obtain progeny: 6, to obtain a son: 9, to obtain a thousand followers 
(anucara) or warriors (vira): 4, to obtain speech (vac): 4, to obtain intelligence 
(medhd) or knowledge (jnana, vidyd): 7, to obtain fame (yasas, varcas): 4, to 
become handsome (rupavat): 1, to obtain success/supernatural power (sid- 
dhi): 6, 8 to obtain concentration of mind (samddhi), sacrifice (makha), truth 

statement of effect. 

6 Rgvidh 3.38d = Rgvidh 3.76d (yamaya saganayaiva) tadbhayam na sa vindati "One 
(who offers to Yama and his retinue) does not have fear from them." See also 3-2-1 
absence of fear (abhaya). 

7 Rgvidh 1.108cd mumursur api ... sarvam ayur avapnuyat "Even one who is about 
to die may obtain a full life span." Rgvidh 2.40b apamrtyum vyapohati "He expels 
death." Rgvidh 1.130cd chittvd sarvan mrtyupasan jived ... "After cutting off all 
nooses of death he may live." Rgvidh 1.1G1& purne mdsijayen mrtyum "After one 
month he may conquer death." Rgvidh 3.13d ghoram mrtyubhayam jayet "He may 
conquer a dreadful fear of death." Rgvidh 3.42c evam yukto jayen mrtyum "Thus 
practicing he may conquer death." For mryumjaya, see Einoo 2005. 

8 In the following four cases the word siddhi can be interpreted as success: Rgvidh 
1.157a, b ... labhate ... siddhim anuttamdm 'he obtains unsurpassed success'; 
Rgvidh 2A3dparam siddhim avapnuyat 'he may reach the highest success'; Rgvidh 
2.106d param siddhim niyacchati 'he secures the highest success'; Rgvidh 2.167d 
arthasiddhih para bhavet 'there may be the highest accomplishment of object'. 
Rgvidh 2.57c svadehe sidhyate jantuh 'a person attains perfection in his very 
body (Bhat 1987: 324).' For a similar idea, see Amoghapasakalparaja 21a, 3 and 
Manjusrimulakalpa 55 [691,3-61. On the other hand, Rgvidh 3.47cd-48 'He sees 
the divine gandharvas, he sees siddhas and caranas. He hides himself from this 
world. He becomes one who goes in the air. He sees from a distance, he listens from 
a distance, like Paramesthin' clearly points to the supernatural powers. 


From kamas to siddhis 

(sattva): 3 
2-1 rites for rain: 5 
2-2 domestic rules: 3 

Rules for taking food: 2, levirate marriage 1 
2-3 rites for agriculture: 4 
2-4 rites for pasturage: 11 
2-5 grace or intervention of the gods: 9 9 

3-1-0 peacefulness (svastyayana): 9 

3-1-1 safety on a journey: 18 

3-1-2 safety when one crosses a river: 4 

3-2-0 pacification (santi): 12 

Pacification of women, men and cows (nannaragosanti): 1, when a dove sits 
down in one's house: 1, pacification of the house: 2, of the planets: 1, of a 
horse: 1, of bad dreams: 6, to calm dissension among relatives (jnatibheda): 
1, to calm down hatred: 1 

3-2-1 absence from fear (abhaya): 4 

3-2-2 for one who wishes purification (suddhikama), or for purification: 3 

3-2-3 release from various evil conditions: 42 

For one who is accused (abhisasta): 3, release from Varuna's noose: 2, one who 
is bound (baddha): 2, release from falsehood (anrta): 1, release from bad luck 
(alaksmi, asrl): 4, from misfortune (duhkha), from various evil conditions 
such as amhas: 1, apad: 3, kalmasa: 1, kilbisa: 1, krcchra: l,papa: 9, rapas: 
1, ripra: 3, other cases: 2 10 , from raksas: 7 

3-3 atonement for one's sins (prayascitta): 19 

9 1 assign Rgvidh 2.54, 64, 65, 78-79, 92cd-95, 179cd-180ab, 180cd-181ab, 3.5cd-6, 
34-37, 4.130cd-131ab to the item 'grace or intervention of gods'. In Rgvidh 2.54d sa 
tasya varado bhavati 'the deity becomes a boon-giver to him,' and Rgvidh 2.92cdya 
icched varadam devim sriyam nityam kule sthitdm 'one who wishes that the boon- 
giving goddess Sri will always stay in one's family' the word varada appears. Ac- 
cording to Bbhtlingk-Roth's Sanskrit Worterbuch, it is only in TA 10.34 and SvetUp 
4.11 that the word varada otherwise occurs. From this we can surmise the idea of 
being given a boon by a god belongs to the latest period of Vedic literature. 

10 Rgvidh 1.118cd-119ab rajakarye svayiithe va abhisasto 'py anekadha 11181 asakye 
pratibhakarye bhaye pranantike 'pi va 'Either in royal service, or among a troop 
of dogs, even when one has repeatedly been infamous, (when engaged) in a deed 
of audacity which cannot be executed, or even in peril of life' (Gonda 1951: 28) is 
classified as 'other cases . Rgvidh 2.121ab yasya nastam bhavet kin cid dravyam 
gaur dvipadam dhanam 'When something like goods, a cow, a man or riches go lost 
for one' is also grouped into this category, because it deals with the bad situation 
that one loses something. The Rgvidh here uses RV 6.54 to find lost things and 
AsvGS 3.7.9 prescribes a rite using the same sukta to find lost things, too. See also 
KausS 37.4-6, 52.12-14 for rites for the same purpose. 


Genesis and Development of Tantrism 

4 rites for women: 12 

To get a maiden (kanya): 1, to get a husband: 1, marriage: 1, conception: 1, 
to beget a son: 1, to abort a dead fetus: 1, to prevent a miscarriage: 2, for a 
safe delivery: 1, to cause to sleep: 1, to expel another wife of the husband: 1, 
to increase virility: 1 

5-1 royal rites: 9 

Royal consecration (rajabhiseka): 1, to obtain kingship (rajya, aisvarya): 5, 
taking the bath of a king: 1, to make an amulet for a king: 1, pacification of 
royal paraphernalia: 1 

5-2 rites for the battle: 8 

6 victory over a rival: 17 

To win a dice game: 1, to win a dispute: 1, to defeat the enemy (satru, dvisant, 
sapatna, ari, ripu): 15 

7 sorcery (abhicara): 26 

To make someone sick: 1, to burn: 1, to expel someone from the country: 1, 
to cause dissension: 3, 11 to control someone through one's will: 9, 12 to kill: 6, 
counter magic: 5 13 

8-1 blessed condition after death/liberation: 21 14 

The abhicaras mentioned hitherto are described in Rgvidh 2.48cd; 1.79-82; 2.49ab; 
1.100, 102-203ab, 2.49cd. 

Among the abhicaras of this kind, Rgvidh 3.84cd-87 and 3.108-109 aim at one's 
kindred or friends. In Rgvidh 3.100cd-107 an effigy of boiled rice is made and this 
sorcery is performed to attract women; it is noted that this rite should not be done 
in relation to a married woman, a female ascetic (sadhvl) or a chaste woman (dhar- 
mavrataslla). According to Rgvidh 2.48ab all things (sarvam) are brought under 
one's control, Rgvidh 2.182cd-183ab claims to subdue the world (jagat) and the sor- 
cery described in 3.95cd-96ab has a king or a country or a town as its victim. Rgvidh 
3.79-84ab prescribes a rite in which an effigy made of clay is used and according 
to the number of days of the performance, either a rich merchant or a ksatriya or 
a king or a brahmin or an ascetic is brought under one's control. In 3.88-92ab an 
effigy made of clay mixed either with ghee or sesame oil or mustard oil is used to 
subdue a brahmin, a ksatriya, and a merchant respectively. According to Rgvidh 
2.42, after one hundred thousand offerings of ashes into water, the sorcery (krtya) 
itself appears from the water. 

Rites to kill an enemy are prescribed in Rgvidh 1.105, 2.16-20, 47, 63, 176cd-177ab, 
3.92cd-95ab. Among them, an effigy of the enemy is used in 2.16-20 and 3.92cd- 
95ab. Also in the battle rite in Rgvidh 2.87cd-89ab, an effigy made of clay is used. 
Examples of counter magic are found in 4.30-33, 36-37, 40-42, 115-116. 
The item concerning a blessed condition after death/liberation is based on descrip- 
tions in the Rgvidh 3.17cd-18ab, 1.106, 3.70cd, 1.107, 136-137, 2.44, 3.18cd-19ab 
(this passage is grouped into two subcategories: to obtain the highest place and 
to obtain immortality), 4.127cd-129ab, l.lllcd-112, 4.19cd-20ab, 1.158d-159a, 


From kamas to siddhis 

To reach the desired goal (ista gati): 1, to obtain heaven (svarga): 2, to ob- 
tain the highest place (para dhaman, para sthana): 4, to obtain immortality 
(amrtatva): 2, to obtain the same realm (salokata) with the moon: 1, to be 
present with all the gods: 1, to go to the world of Indra: 1, to obtain intimate 
union (sayojya) with Prajapati: 1, to go to the highest place of Visnu: 1, to 
go to the world of Brahman: 3, to obtain brahmanirvana: 1, not to be born 
again: 2, to obtain memory of previous lives (jatismara): 1 
8-2 blessed condition of the ancestors: 3 

The results of rites the performer undertakes are generally classified as follows: 
he maintains a good condition and seeks for a better condition, but when a bad 
condition happens, he tries to recover from it. Among the items given above, the 
maintaining of a good condition corresponds to 3-1 peacefulness (svastyayana). 
The seeking of a better condition is, generally speaking, item 2 prosperity 
(paustika), and when the better condition is concerned with life, it is a matter 
of item 1-1 to live a full life span (ayusya). Recovering from a bad condition is 
3-2 pacification (santi), and if the bad condition is disease, item 1-2 remedy for 
a disease deals with it, while in the case of a transgression of social rules, 3-3 
atonement for one's sins (prayascitta) is measured against it. Item 6 victory over 
a rival can be, broadly speaking, likened to recovering from a bad condition, the 
rival being representative of the bad condition. 

M. Bloomfield classifies the ritual hymns into nine classes: 1) Charms to 
cure diseases and possession by demons (bhaisajyani), 2) Prayers for long life 
and health (ayusyani), 3) Imprecations against demons, sorcerers, and ene- 
mies (abhicarikani and krtyapratiharanani), 4) Charms pertaining to women 
(strikarmani), 5) Charms to secure harmony, influence in the assembly, and 
the like (sammanasyani, etc.), 6) Charms pertaining to royalty (rajakarmani), 
7) Prayers and imprecations in the interest of Brahmans, 8) Charms to secure 
prosperity and freedom from danger (paustikani), and 9) Charms in expiation 
of sin and defilement (prayascittani) (Bloomfield 1899: 57). Taking Bloomfield's 
classification into consideration, I add further item 4 rites for women, item 5 

4.44cd-45ab, 3.137, 2.67, 3.2cd-5ab, 3.75. Brahmanirvana is mentioned as one of 
the results of a very long rite for the worship of and meditation on Visnu Narayana 
by using the Purusasukta (RV 10.90): 3.149cd munayah sarve brahmanirvanam 
dpnuyuh 'all sages may reach the brahmanirvana' . There are two rites so that one 
will not be born again: a short version in 2.186cd and a long one in 4.20cd-28, 
the latter using the Ratrisukta (RV 10.127) which corresponds to Samavidh 3.8.1-5 
(Gonda 1951: 103). The result of remembering previous lives is mentioned in Rgvidh 
2.45cd, and Samavidh 3.7.1 also promises the same result. See also Rgvidh 3.36cd. 
The next item 8-2 'blessed condition of the ancestors' is based on the descriptions in 
Rgvidh 1.109, 3.8ab, 4.129cd-130ab. 


Genesis and Development of Tantrism 

royal rites, and item 7 for sorcery (abhicdra). Item 8, the blessed condition after 
death, is a new article I have introduced for the Rgvidhana. 

3 Analysis of the rites of the Samavidhanabrahmana 

The Samavidhanabrahmana 15 has three chapters. At first glance the contents 
of the text seem to be arranged systematically Samavidh 1.1.1-7 describes the 
creation of the world, 1.1.8-18 praises the Vedic melodies (samans), 1.2.1-12 
prescribes three kinds of penances (krcchra) and 1.3-4 is dedicated to the study 
of the samans (svadhyaya). Then in 1.5-8 various kinds of atonements for dif- 
ferent sins are discussed, through to the end of the first chapter. The second 
and third chapters are divided into sections introduced by the phrase 'and from 
now' (athatah) and the contents are summarized, for example, 2.1.1 athatah 
kamyanam II 2.1.3 ayusyany eva prathamam II "And from now on the rites 
for special desires. At first rites for longevity (will be prescribed)." 16 In the de- 
scription of the section on rites for longevity there are naturally prescribed many 
rites for this purpose as well as similar rites 17 but we can also find other types of 
rites there as well. 18 The situation is also the same with other sections dealing 
with other kinds of rites. 

Now follows an analysis of rites in the Samavidhanabrahmana; 19 the scheme 
of analysis is the same as that used for the Rgvidhana. 

1-1-0 to live a full life span (ayusya): 4 

To live long: 2.1.10; 11, to live for a hundred years: 2.1.5; 2.2.1 
1-1-2 to conquer death: 3 

15 I use the following edition of the Samavidhanabrahmana: Sdmavidhdna Brdhmana 
with Veddrthaprakdsa of Sayana and Paddrthamdtravivrti of Bharatasvamin, crit- 
ically edited by Dr. B. R. Sharma, Tirupati: Kendriya Sanskrit Vidyapeetha, 1964. 

16 See further Samavidh 2.6.1 athatah saubhagyanam (on the rites for nuptial hap- 
piness), 2.6.17 athdto yasasydnam (on the rites for fame), 2.7.1 athdto brah- 
mavarcasydndm (on the rites for holy luster), 2.8.1 athatah putriydndm (on the rites 
for obtaining sons), 3.1.1 athdto dhanydndm (on the rites for becoming wealthy), 
3.3.6 athdto vdstusamanam (on the pacification of the dwelling place), 3.4.1 athdto 
drstadarsandndm (on the rites to see invisible things). 

17 For example, Samavidh 2.1.5 and 2.2.1 rites to live for one hundred years, 2.1.10, 
11 rites to live long, 2.4.9 a rite to ward off old age and death, 2.3.7, 8, 10, 11 rites 
not to die accidentally. 

18 For example, Samavidh 2.1.7 an atonement for when one receives too much, 2.4.8 
blessings when going to sleep and when getting up, 2.5.1-4, 6-7 rites for the subju- 
gation of various beings, 2.5.5 a rite for killing one's enemy. 

19 As the number of cases in this text is much less in comparison with those in the 
Rgvidhana, the place in the text is given for each case. Some items mentioned in 
the analysis of the Rgvidh are skipped due to the lack of passages dealing with 


From kamas to siddhis 

Not to die of thirst: 2.3.7, not to die in the water: 2.3.8, to expel decrepitude 

and death: 2.4.9 
1-2-0 remedy for disease (bhaisajya): 3 

When one becomes ill: 2.2.3, when a part of the body aches: 2.3.1; 2 
1-2-1 remedy for consumption (yaksman): 1 (2.3.9) 
1-2-3 remedy for poison: 2 (2.3.10; 11) 

2-0 prosperity (paustika) or for various desires (kamya): 47 

To obtain wealth (sn): 3.1.3; 4; 3.2.6, to obtain a thousand (sahasra): 3.1.10; 
11; 13; 3.2.1, to become wealthy (dhanya): 3.1.2, to obtain food: 2.3.5; 6, to 
obtain gold: 3.1.8; 3.3.3, to obtain gold or silver: 3.1.12, to obtain a village: 

3.2.4, to obtain a son who is handsome and long-lived: 2.8.2; 3; 4; 5, to obtain 
a hundred retinues: 2.8.6, to obtain fame: 2.6.16; 18, to obtain glory of the 
brahmin: 2.7.1; 3, to obtain trust: 2.7.2, to speak of what has been heard 
(srutanigadin): 2.7.4-11, to be good at talking (kathdsu sreyas): 2.7.12, to 
speak against the king: 2.7.13, 20 to become fortunate: 2.6.2-5, to be liked by 
all: 2.6.6, to obtain divine prosperity (daiva posa): 3.3.4, to obtain prosperity 
of the asuras (dsura posa): 3.3.5, to obtain all human desires (manusa kdma): 
3.9.3, to obtain all divine desires (daiva kdma): 3.9.4, to obtain the position 
of the lord of the three worlds: 3.9.5, a rite in which, by each saman of a 
set of eight samans, one obtains gold, corn, cattle, son, village, fame, glory 
of the brahmin, and heaven respectively: 3.2.5, a rite in which according to 
the depth of water in which one sinks to sing a saman, one obtains wealth 
(laksmi), corn, cattle, son, village, and a thousand respectively: 3.2. 7-12. 21 
Supernatural powers (siddhi) 

To see what is hidden: 3.4.2-5, divination of success (siddhi): 3.4.6; 7; 9, 
divination of a good crop: 3.4.8, divination of a victory: 3.4.10, divination of 
longevity: 3.4.11, to see the ancestors: 3.7.4, to see gandharvas and apsaras: 

3.7.5, to see the gods: 3.7.6, to find a hidden treasure (nidhi): 3.7.7-8, to see 
demons (bhuta) and receive money from them: 3.7.9, to walk in the air: 3.9.1, 
to move as swift as thought: 3.9.2 

2-3 rites for agriculture: 2: to obtain crops: 3.1.9, 3.2.3 

20 We find similar results in the Hemasadhanapatala: Manjusrimulakalpa 55 
[697.10-11] rajakule cottaravadi bhavati 'and he becomes one who defends himself 
in the royal house,' [690.16-17] rajakulesuttaravadi bhavati. See further [679.13- 
15], [695.2-3], [690.17-18], [690.15-16],'[680.11-14], [695.2-3], [699.14], [707.25- 
26], [710.8-9], [719.15-16]. 

21 We can find specifications about how deep one sinks in water: for example in 
GobhGS 4.5.26, AVPS 36.26.1-3, Manjusrimulakalpa 55 [672.7-10], [672.22-24], 
[673.19-21], [678.24-26], [691.8-9], Amoghapasakalparaja 44b, 1-2 [59.17-23]. 


Genesis and Development of Tantrism 

2-4 rites for pasturage: 3: to obtain cows: 3.2.2; 3.3.1, to obtain a number of 
cattle: 3.3.2 

3-1-0 peacefulness (svastyayana): blessings at the time of sleeping and waking: 

3-1-1 safety on a journey: 2.4.1-7 
3-2-0 pacification (santi): 2 

Pacification of the house: 3.3.6-7, pacification of omens: 3.5.5 22 
3-2-1 absence of fear (abhaya): 2 

To avert fear of snakes: 2.3.3, to avert fears from weapons: 2.3.4 
3-2-2 purification: to become purified (puta): 2.1.6 
3-2-3 release from various evil conditions: 5 

From bad luck (alaksmi): 3.1.5; 6; 7, from obstruction (sambadha): 2.1.9, 

from raksas: 2.2.2 

4 rites for women: 2 

Taking a bath before sexual intercourse: 2.6.12, to wed one's daughter: 

5-1 royal rites: royal consecration (rajabhiseka): 3.5.1-4 
5-2 rites for the battle: 3.6.1-7; 8; 9; 11 

7 sorcery (abhicara): 

To burn: 3.7.2, to allure a woman: 2.6.7; 11, to subdue a woman: 2.6.8; 9; 10, 
to attract courtesans (vesastha) and female ascetics (pravrajika): 2.6.13, to 
win over either a brahmin or a ksatriya or a vaisya or a sudra: 2.5.1-4, to win 
over various beings according to a period of the performance of the rite (for 
one night: the family, for two nights: retinues of the king, for three nights: 
the king, for four nights: a village, for five nights: a town, for six nights: a 
country, for seven nights: asuras and raksas, for eight nights: the ancestors 
and pisacas, for nine nights: yaksas, for ten nights: gandharvas and apsaras, 
for a half month: Vaisravana, for a month: Indra, for four months: Prajapati, 
for a year: the whole world): 2.5.2, to subdue pisacas: 3.7 .3, jambhakas grant 
all desires: 3.7.10, to kill: 3.6.10; 12; 13 
counter-sorcery: 3.5.6; 7; 8 

22 In Samavidh 1.5-8 various kinds of atonements for different sins are discussed and 
at the end of the passages the pacifications of some omens are described (Gampert 
1939: 5): 1.8.7: bad dreams, 1.8.8; 12: some indefinite omens, 1.8.9: fire burning, 
1.8.10: damage caused by mice, 1.8.11: when a seat (kurca) is broken, 1.8.13-15 
describes pacifications of various disasters (abhivata) among men or cows or horses. 


From kamas to siddhis 

8-1 blessed condition after death/liberation 

To go to the world of Brahma: 2.1.8; not to be born again: 3.8.1-5, to obtain 
the memory of previous lives (jatismara): 3.7.1 

When we compare the items of the Rgvidhana and the Samavidhana, item 2-0 
'prosperity (paustika) or various desires (kamya)' is the most numerous in the 
Rgvidhana. As regards the siddhis or supernatural powers, the Rgvidhana gives 
only one example, namely Rgvidh 3.47cd-48, which I refer to in note 8. Pas- 
sages dealing with the supernatural powers in the Samavidhana are collected at 
the end of item 2-0, where the examples amount to fourteen in number. Rites 
to obtain siddhis in the Samavidhana are therefore much greater in number 
than in the Rgvidhana. The same thing can be said about the rites for sorcery 
that comprise item number 7. I have listed 26 rites for the Rgvidhana, but only 
17 for the Samavidhana. But in terms of percentage, those in the Rgvidhana 
amount to only 6.3 percent, while those in the Samavidhana amount to 17 per- 
cent. Item number 2-5 'grace or intervention of the gods' does not appear in the 
Samavidhana, while item number 3-2-3 'release from various evil conditions' and 
item number 8-1 'blessed condition after death' clearly show a decrease in that 
text. From this we can say that the Samavidhana shows more interest in things 
supernatural and magical while at the same time paying sufficient attention to 
mundane matters. 

4 Analysis of the rites of the Gobhilagrhyasutra (4.5-9) 23 

Now I turn to an analysis of the Gobhilagrhyasutra (4.5-9), which belongs to 
the tradition of the Samaveda. The Gobhilagrhyasutra, book 4, chapters 5 to 
9 provides a comparatively systematic description of rites performed for the ob- 
tainment of different desires. As the opening sutra GobhGS 4.5.1 suggests that 
the rites which follow are clearly defined as kamya (kamyesu ata urdhvam 'in 
the rites for the obtainment of definite desires, which will be described from now 
on'). The result of the analysis is as follows. 

1-1-0 to live a full life span (ayusya): for one who desires a full life span: 4.8.11- 

1-1-2 to conquer death: to avoid an undesirable death: 4.6.1 

23 The Khadiragrhyasutra (4.1.1-4.4.4) describes corresponding rites to fulfill var- 
ious desires, but the analysis is based mainly on the description of the 
Gobhilagrhyasutra. As GobhGS 4.7 is a detailed description of the construction 
of a house, this chapter is left out of consideration. 


Genesis and Development of Tantrism 

1-2-0 remedy for disease (bhaisajya): 24 

To make clean the place where there are worms (krimi): 25 4.9.18 
1-2-3 remedy for poison: for one who has been bitten by a venomous animal: 26 


2-0 prosperity (paustika) or for various desires (kamya): 

For one who desires enjoyment (bhogakama): 4.5.28, for one who desires to 
obtain a hundred cart-loads or a thousand cart-loads: 4.6.13; 4.9.11, for one 
who desires that his means of livelihood may not be exhausted: 4.8.19; 4.9.11, 
offering of a part of the articles of trade (panyahoma): 27 4.8.20-22, to obtain 
property on the earth (parthiva): 28 4.5.21-26, to obtain a village: 4.8.14-18, 
for one who desires sons and cattle: 4.5.15-16, for one who desires compan- 
ions (sahayakama): 4.8.23-25, for one who desires the glory of a brahmin: 
4.5.14, for one who desires fame: 4.6.10; 4.8.23-25, to obtain the favor of a 
person: 4.5.18-20; 4.8.8-10 

2-4 rites for pasturage: 29 

For one who desires cattle: 4.5.15-16; 4.9.6; 12; 13, for one who desires safety 
of cattle: 4.5.17; 29; 30, against worms of cattle: 4.9.19-20, when the cow- 
stable is burnt: 4.9.7 30 

3-1-0 peacefulness (svastyayana): 31 

A baliharana is performed at the end of the sravana and agrahayani 32 at a 
cross-road to the northeast of the village for the sake of svastyayana: 4.8.1-7, 
for the svastyayana of the snataka: 4.9.16-17 

25 Besides KhadGS 4.4.3-4, KausS 27.14-20 and KausS 29.20-26 describe rites to 
cure worms (Bloomfield 1899: 61). 

26 p or r it es t cur e snakebite and similar accidents, see KhadGS 4.4.1, and KausS 
29.1-14, 29.28-29, 31.26, 32.5-7, 32.20-25. SankhSS 16.13.3-4 is counted in this 
group (Zysk 1992: 105). 

27 See also KhadGS 4.3.7. ApGS 8.23.5; HirGS; KausS 41.8-9; 50.12- 
16; 59.6 describe rites to pray for success in business. AsvGS 3.7.8; GobhGS 4.5.33; 
ApGS 3.9.2; KausS 42.1-5; KausS 50.12-16 are rites for success of a journey for 

28 Bhattanarayana, a commentator, comments on GobhGS 4.5.22 as follows: 
prthivyartham kriyata iti parthivam gramaksetradilabhartham 'In that it is per- 
formed for the sake of the earth it is called parthiva, namely in order to obtain a 
village, field and so on.' 

30 GobhGS 4.9.6-7 gosthe pasukamah 16/ viduyamane civaram 111%. One who desires 
cattle (performs it) in the cow-stable. 7. If (the cow-stable) is damaged by fire, (he 
should offer) a monk's robe.' For the interpretation of this rite, see Knauer 1884: 
208 and Oldenberg 1892: 127. 

32 GobhGS 3.7.1-23 and GobhGS 3.9.1-21 describe the sravana and agrahayani re- 


From kamas to siddhis 

3-1-1 safety on a journey: 33 4.5.31; 32; 4.9.8-10 

3-2-1 absence of fear (abhaya): 34 from serious diseases and sorcery: 4.6.2 
3-2-3 release from various evil conditions: to expel misfortune (alaksmT): 35 

5-1 royal rites: for one who desires the position of the lord of the people (puru- 
sddhipatyakdma): 4.9.1-5 

7 sorcery (abhicara): for one who desires killing (vadhakama): 4.8.11-13 

As the opening sutra suggests, the rites described here are kdmya; there- 
fore, among the items of the Gobhilagrhyasutra item number 2-0 'prosperity 
(paustika) or for various desires (kamya)' is the most numerous. Desires 
concerning fields, villages, sons and cattle, the prestige of a brahmin and fame 
are regarded as the traditional topics of the Vedic kamyestis. A rite to obtain 
the favor of a person can be counted as an Atharvavedic interest in securing 
harmony (sammanasya) (Bloomfield 1899: 72-73). The rite of offering a part 
of the articles of trade for the sake of success in business is similar to rites in 
other Grhyasutras, as pointed out in note 27. It is remarkable that rites for long 
life and to remedy disease, which appeared in great numbers in the Rgvidhana 
can hardly be found in the Gobhilagrhyasutra. The reason that the number of 
rites for long life is so low may be due to the fact that many of the rites of the 
Grhyasutras, especially those performed at the different stages of life, such as 
the rite of a new-born child, have as their very important aim the prayer for the 
long life of a growing child. The Grhyasutras indeed have many occasions to 
pray for the long life of a person, so it was not necessary to collect rites for this 
purpose in particular (Bloomfield 1899: 64). 

Rites to remedy disease are collectively described in the Kausikasutra 25.1- 
32.27 and it seems as if medical rites were the monopoly of the Atharvavedic 
tradition. Only a limited number of rites for remedies were treated in other 
Grhyasutras proper (see note 24). It may be natural therefore that the number of 
such rites in the Gobhilagrhyasutra should be very few. Even the Rgvidhana and 
the Samavidhanabrahmana describe a few agricultural rites, so it is remarkable 
that the Gobhilagrhyasutra does not mention any at all here, but treats them 
in another place, namely in GobhGS 4.4.27-34. The Gobhilagrhyasutra, on the 
other hand, enjoins some rites for pasturage or cattle breeding in the collection 
of kdmya rites, but, as in the case of agricultural rites, this Grhyasutra also pre- 
scribes rites for pasturage in another context, namely, in GobhGS 3.6.1-15. The 
svastyayana in general, and especially safety on a journey, were matters of con- 
cern from Vedic times. In item number 3-2-1, namely absence of fear, I mention 


Genesis and Development of Tantrism 

absence of fear from serious diseases and sorcery. Absence of fear from serious 
diseases can be grouped into the item ayusya and the latter can be classified un- 
der "sorcery", because it is a kind of a counter magic. There are also some hymns 
and rites in the Vedic texts regarding absence from fear (see note 34), thus this 
item is again traditionally Vedic. As for item number 3-2-3, a rite to expel misfor- 
tune, I could find only one example in a sister sutra, Khadiragrhyasutra 4.1.20- 
21, but it seems that during the marriage ceremony, a certain rite was performed 
to expel alaksmT from the bride, as I mention in note 35. 

I would thus like to maintain that the kdmya rites of the Gobhilagrhyasutra, 
a Vedic text, aim at attaining the various desires well-known in the Vedic texts. 
The items of the Rgvidhana vary widely but most can be assumed to belong to 
the traditional objectives of desires. Among the items in the Rgvidhana, the fol- 
lowing may indicate a new tendency. As I mentioned in note 9 concerning item 
number 2-5, namely grace or intervention of the gods, the gods become varada or 
boon-givers. As the examples given in note 12 on item 7 sorcery show, the various 
rites to control someone are almost the same as rites of subjugation (vaslkarana) 
that appear in the tantric six acts (satkarmani). In notes 12 and 13, 1 referred to 
several cases in which an effigy is used in the rite of sorcery. Even though it has 
already been pointed out that the use of an effigy in a magical act was already 
known in the Atharvaveda (Gonda 1980: 255), and the Kausikasutra gives sev- 
eral cases of it (see KausS 35.28, 36.14, 47.54), the usage of the effigy in sorcery 
tends to be more popular in tantric magic; according to Bbhtlingk and Roth's 
Sanskrit Worterbuch the word jdtismara or remembering of one's former life ap- 
pears from the Mahabharata onward. There is only one case in the Rgvidhana 
(see note 14), but this may represent a new tendency. 

The Samavidhanabrahmana also generally gives the traditional desires, but 
the following indicate new tendencies. Mentioned among the various desires 
is divine prosperity (daiva posa), prosperity of the asuras and divine desires 
(daiva kdma), as is obtaining the position of lord of the three worlds. Most no- 
table though are the statements enumerated under the headings of supernatural 
power: seeing what is hidden seems to be peculiar to the Samavidhana. I have 
mentioned several kinds of divination. Divination itself was already known in 
the Vedic texts (Thite 1978) and the examples given there are not something 
new, but the ensuing examples of siddhi, namely to see ancestors, gandharvas 
and apsaras, gods, demons and hidden treasure (nidhi) have so far been found 
in the broadest sense among Vedic texts only in the Samavidhanabrahmana (see 
item 2.0), Rgvidhana (see note 8) and Atharvavedaparisista. Rites of sorcery are 
mentioned in the Samavidhanabrahmana (item number seven, sorcery) and this 
too is an indication of the new tendency. As I have mentioned in note 12, the 


From kamas to siddhis 

targets of subjugation appearing in the Rgvidhana are at most king and coun- 
try, but in the Samavidhanabrahmana, supernatural beings such as demons and 
gods are brought under the control of the practitioner. 

The Atharvavedaparisista, chapter 35 asurikalpa (Magoun 1899) and chap- 
ter 36 ucchusmakalpa (Bisschop, P. and A. Griffiths 2007), describe a number 
of rites of subjugation and Atharvavedaparisista 36.2.9 claims to subjugate even 
the Isvara. According to AVPS 36.25.1-4, an excellent woman or a woman grant- 
ing desires (varastri) appears as the result of the rite and grants whatever the 
performer desires. Thus, in texts complementary to the Grhyasutras such as 
the Rgvidhana, Samavidhana and Atharvavedaparisista we can find many new 
types of the rites clearly exhibiting magical elements, and those rites are pre- 
sumed to link with tantric rites. 

5 Analysis of the Vinasikhatantra (151-224 and 264-300) 

The Vinasikhatantra 36 is one of the early Saivatantras from the north of India 
(Goudriaan 1985: 4). In this text a deity called Tumburu and his four saktis or 
sisters, Jaya, Vijaya, Ajita and Aparajita, play a central role. 37 Vinasikhatantra 
151-136ab teaches their bijamantras 38 and Vinasikhatantra 151-224 and 264- 
300 prescribes a variety of mainly magical rites, whose analysis follows: 

1-2-0 remedy for disease (bhaisajya): 183-184ab; 184cd 

2-0 prosperity (paustika) or for various desires (kamya): 

What one uses becomes inexhaustible: 185, 39 to be liked by all: 186, 40 for 
one who desires welfare (snkama): 187-188ab, for one who desires all (sar- 
vakama): 41 188cd-1901b; 218-224ab; 289cd-291, an immediate elevation of 

36 I use the following text: T. Goudriaan, The Vinasikhatantra, a Saiva Tantra of the 
Left Current, edited with an Introduction and Translation, Delhi: Motilal Banarsi- 
dass, 1985. 

37 For Tumburu and his four sisters, see Goudriaan 1973 and Goudriaan 1985: 18-30, 
47-48. For the fact that Tumburu was known in South East Asia, see Goudriaan 
1985: 24 and Sanderson 2001: 8 and for the fact that he was also known in Chinese 
Esoteric Buddhism, see Sanderson 2001: 8. 

38 For their different bijamantas, see Goudriaan 1973: 50. 

39 We find similar rites in GobhGS 4.8.19 and 4.9.14 for one who desires that his means 
of livelihood may not be exhausted. 

40 A similar rite is found in Samavidh 2.6.6. AV 19.62.1 conveys a similar idea. 
This kind of rite seemed to be very popular among the Buddhists. See, for ex- 
ample, Amoghapasakalparaja 33a,l and the following places in the Hemasadhana: 
Manjusrimulakalpa 55 [714.26-271; [671.28]; [673.14-15]; [686.25-28]; [698.18- 
19]; [707.4]; [714.24-26]; [716.2]; [716.5-6]. 

41 Many passages among the twenty passages listed at the beginning of item 2-0 of 
the Rgvidhana, namely 'to obtain desires (kama)' are for one who desires all (sar- 


Genesis and Development of Tantrism 

one's position (sadyotkarsana): 42 162 

3-2-0 pacification (santi): santi and pusti are reached: 181-182 

4 rites for women: 

To make someone impotent: 277-278, to increase virility: 282-283ab 

7 sorcery 

To attract a woman: 151-154; 194-196, to attract someone: 201cd-206, 
266cd-268ab, to revive a dead person who grants a desire: 190cd-193, 43 
how to prepare a magic ointment (ahjana) to make someone as charming 
as the god of love: 279-281, 44 to bring someone under control {vasikarana): 
268cd-269; 274cd-276; 283cd-286; 286cd-289ab, vasikarana of the enemy: 

vakama) and in order to obtain all desires: Rgvidh 1.144; 124ab; 160cd-161ab; 2.43; 
2.44; 2.45ab; 2.165-166; 2.178cd-179ab; 3.46-48; 3.138-142; 4.7cd; 4.29. KausS 
59.19-20 already describes a rite for a sarvakdma. 

Goudriaan 1985: 39 finds the meaning of utkarsana to be unclear, but on p. 114 
he translates it 'total uprooting [of the enemy]' and considers it to be a synonym 
for uccatana. Vinasikh 162 reads as follows: grhitvd tu mahdmdmsam dadhi- 
madhvdjyasamyutam I dhutydstasahasrena sadyotkarsanam uttamam II 'Having 
collected human flesh together with sour milk, honey and clarified butter, an imme- 
diate elevation of one's position is obtained by offering them a thousand and eight 
times.' A rite called uccatana is known in the Vinasikhatantra, because it uses the 
word in Vinasikh 165a. The association of utkarsana with uccatana may have been 
caused by the use of human flesh (mahdmdmsa); a rite which uses human flesh 
must be of a cruel character. There are another two cases in the Vinasikhatantra 
in which human flesh is used. In Vinasikh 189cd-190ab there is a rite for one who 
desires all, in which human flesh (naramdmsa) and the flesh of goats are offered 
one hundred thousand times. Vinasikh 190cd-193 describes a magical rite to re- 
vive a dead person who then grants one of the performer's desires. As these two 
other cases involving the use of human flesh belong to the rite to fulfill desires, it is 
not unlikely that utkarsana can mean the elevation of one's position. 
As mentioned in note 42 a dead person revives and grants one of the practitioner's 
desires. There are similar rites in other texts. In Rgvidh 2.42, when one offers 
ashes one hundred thousand times in water, the personified sorcery (krtyd) appears 
from the water. According to Samavidh 3.6.12-13, when mustard oil is offered a 
thousand times at a crossroad on the fourteenth night of the waning half month, a 
person having a sword in his hand appears and when he is ordered to kill someone, 
he does so. In AVPS 36.25.1-4 an excellent woman or a woman granting desires 
(varastri) appears and says that she will fulfill whatever the performer desires. See 
Goudriaan 1978: 294-298, for similar rites in the later Tantric texts. 
For the magical ointment applied to the eyes (anjana or anjana), see Gonda 1980: 
150-152 and Goudriaan 1978: 317-318. Goudriaan 1978: 317 refers to AVPS 35.2.2, 
but the following 35.2.2cd-3ab and 35.2.3cd-4ab teach the preparation of two 
kinds of anjana. By applying the first, everybody becomes one's servant (kimkara) 
while the second is called the origin of good fortune of all beings (sarvabhutdndm 
saubhdgyasya tu kdranam). Samavidh 2.6.5 teaches a sdman for the anjana to be- 
come happy. 


From kamas to siddhis 

168-169; 170, of the king: 174-177, of the king or the queen: 159-161, the 
three worlds (trailokya): 45 163-164, 270-274ab, to cause hatred (vidvesana): 
171-173; 197-199ab; 216cd-217, to uproot the enemy (uccatana): 165-167, 
to kill the enemy (marana): 155-158; 178-180; 207-216ab 

In the Vinasikhatantra the variety of rites is extremely limited. Most of them 
are grouped into the rite of sorcery, and some other rites seemingly belong to the 
fantastic imagination such as those prescribed in verses 185 and 186 (see item 2- 
0). The situation is rather similar to the rdjasa and the tdmasa siddhis described 
in the Siddhayogesvarimata mentioned above (p. 18). 

6 Analysis of the Amoghapasakalparaja (43b.6-44a.7) 46 

The Amoghapasakalparaja 43b.6-44a.7 makes up an independent chapter and 
deals with homa rituals (Kimura 2005). The contents of this short chapter are 
analysed as follows: 

2-0 prosperity (paustika) or for various desires (kamya): 

To obtain the highest prosperity: 44b,3-4 [60.2-4], to obtain one hundred 
dinaras: 44a.2-3 [58.6-10]; 44b. 1 [59.13-15], 47 to attract ayaksini who gives 

45 Goudriaan 1978: 299-300 refers to "attraction of the threefold inhabited world" 
(trailokyakarsana) and Buhnemann 2000: 28-29 mentions a cakra which deludes 
the threefold world (trailokyamohanacakra). 

46 Some parts of the Amoghapasakalparaja have been published in the form of tran- 
scribed texts, as follows: T. Kimura, N. Otsuka, T. Sugiki eds., 1998, "Tran- 
scribed Sanskrit Text of the Amoghapasakalparaja, Part I," Annual of the Insti- 
tute for Comprehensive Studies of Buddhism, Taisho University, No. 20, pp. 304— 
251; Y. Ito, R. Kouda, Y. Matsunami eds., 1999, "Transcribed Sanskrit Text of the 
Amoghapasakalparaja, Part II," Annual of the Institute for Comprehensive Studies 
of Buddhism, Taisho University, No. 21, pp. 154-107; A. Suzuki, N. Otsuka, H. 
Kimura eds., 2000, "Transcribed Sanskrit Text of the Amoghapasakalparaja, Part 
III," Annual of the Institute for Comprehensive Studies of Buddhism, Taisho Uni- 
versity, No. 22, pp. 372-309; Y Ito, H. Yaita, S. Maeda eds., 2001, "Transcribed 
Sanskrit Text of the Amoghapasakalparaja, Part IV," Annual of the Institute for 
Comprehensive Studies of Buddhism, Taisho University, No. 23, pp. 406-331. In the 
course of the analysis of the contents of Amoghapasakalparaja 43b.6-44a.7 I some- 
times refer to the Hemasadhanapatala, the last chapter of the Manjusrimulakalpa, 
pp. 668-722 where a great number of esoteric rites are briefly described. For the 
Manjusrimulakalpa I use the following text: The Aryamanjusrimulakalpa , ed. by 
T Ganapati Sastri, 4 parts, Trivandrum, 1925 (reprint, Delhi: Sri Satguru Publica- 
tions, 1989). 

47 AVPS 36.26.1-3 prescribes a rite to obtain one hundred dinaras and one hundred 
cows. In the Amoghapasakalparaja I can find only one rite to obtain one thousand 
dinaras (Amoghapasa 30a. 7). In the Hemasadhana in the Manjusrimulakalpa it 
seems that the obtainment of various amounts of dinaras is a highly favorite pur- 


Genesis and Development of Tantrism 

five thousand rupakas: 44a.3-4 [58.12-16], 48 to obtain one thousand gold: 
44b.l [59.15-17], 49 to obtain a great amount of treasure (ratna) from a mirac- 
ulous girl who appears from a river: 44b. 1-2 [59. 17-23], 50 to obtain one thou- 
sand villages: 44a.2 [58.5-6], 51 to obtain supernatural powers (siddhi), to find 
hidden treasure (nidhana): 44a.6-7 [58.26-59,4], 52 the magical practice of a 

pose. Thirteen rites for obtaining one hundred dinaras are taught in Manjusri 
55 [671.11-12], [[673.2-3], [678.4-5], [684.23-24], [688.2-3], [688.26-27], [688.27- 
689.1], [689.1-2], [706.28-707.2], [702.2-4], [708.23-24], [711.11-12], and [712.28- 
29]. Various amounts of dinaras can be obtained: one dinara in 55 [674.21-22] 
and [679.3-7]; four dinaras in [695.13]; five dinaras in [671.12-13] and [711.9-11]; 
seven dinaras in [707.8-12], twenty-five dinaras in [673.1-2], more than one hun- 
dred dinaras in [671.15-16] and [688.22-26]; three hundred dinara sin [671.14-15]; 
five hundred dinaras in [678.13-15], [694.17-19], [700.3-5] and [711.5-9]; one thou- 
sand dinaras in [671.24-26], [675.2-4], [676.24-25], [676.28-277.2], [684.25-26], 
[685.2-3], [685.6-7], [685.7], [688.9-11], [688.12-20], [709.29-710.2], [711.12-14], 
[712.14], [718.5-6] and [718.20-21]; five thousand dinaras in [677.2-4] and [694.17- 
19]; twelve thousand dinaras in [706.24-27]; one hundred thousand dinaras in 
[671.11], [672.26-27], [685.5] and [707.6-8]. 

48 The Hemasadhana again hands down several rites to obtain many rupakas: 
[673.25-27], [678.27-679.2]: one hundred rupakas; [677.4-5], [678.16-19], [688.4- 
6]: one thousand rupakas; [692.8-10]: seven thousand rupakas. As a yaksini ap- 
pears in this rite, it can be classified as a rite for attraction (akarsana). 

49 A rite to obtain one thousand gold is found in AVPS 35.2.6ab. The Hemasadhana 
teaches similar rites in the following places: [671.22-23], [672.7-10], [672.15-16], 
[684.15-16], [684.18-19], [684.26], etc. Amoghapasa 44b,l [59.16-17] says that af- 
ter obtaining gold one should worship the three treasures, distribute some part of 
it, and take the rest of it. In Amoghapasa 44a, 6-7 [58.26-59.4] a rite to obtain hid- 
den treasure is prescribed and at the end of the rite it enjoins that one should give 
the proper share to the tree treasures and take the rest of it. See also Amoghapasa 
44b. 2 [59.22-23]. In the Hemasadhana a similar idea is expressed by the word 
ratratrayopayojya in [676.6-7], [678.7-8], [678.15], [678.19], [677.28]. 

50 As in this case a miraculous girl appears from a river, we can classify it as akarsana. 
Incidentally, this rite is performed on the bank of a river which flows down to the sea 
(samudragamini nadi). According to the rite in Amoghapasa 44a. 4-6 [58.16-26] a 
yaksini appears and becomes the female servant of the practitioner, and this rite too 
is performed on the bank of a river running to the sea. The Hemasadhana describes 
several rites which are to be performed at the same place: [669.15], [672.8], [672.10], 
[672.22], [674.7], [674.16], [677.8], [678.6], [683.18], [686.13]. 

51 I do not know of any other rites to obtain as many as one thousand villages. In many 
cases the village is singular and one obtains one village. The villages that can be ob- 
tained by a rite in Samavidh 3.2.5 are plural but the number is not expressed. When 
the number of the villages is expressed, it is rather moderate; KausS 22.7 describes 
a rite to obtain seven villages. According to GobhGS 4.8.15-18 one gets seven vil- 
lages, if one succeeds, but one can get at least three villages. The Hemasadhana 
also teaches several rites to obtain various numbers of villages: three villages in 
[716.18], five villages in [677.9-11], [712.14-15], and [718.13-14], eight villages in 
[696.3], ten villages in [672.11-13], and twelve villages in [704.25-27]. For the idea 
in the Vedic texts to wish to obtain a village, see Kulke 1991. 

52 As Goudriaan 1978: 307 has already mentioned, Samavidh 3.7.7 and 8 and AVPS 
35.2.4 enjoin similar rites. As regards the Manjusrimulakalpa, T Goudriaan says 


From kamas to siddhis 

cave (bilasadhana): 44b.5 [60.13], to open all doors of a cave: 44b.5 [60.14], 
to enter a forest and to attract all goods (dravya) and medical herbs: 44b. 5-6 
[60.15-17] 53 

2-1 rite for rain: 44b.4 [60.6-8] 54 

3-2-0 pacification (santi): 

Protection of the house (grharaksa): 55 43b.7-44a.l [57.15-26], great pacifi- 
cation (mahasanti): 44a.l-2 [57.26-58.1]; 44b,3 [60.1-2], to stop excessive 

that in this text this kind of rite appears repeatedly and he refers, for example, to 
Vol. II, p. 299 and Vol. Ill, p. 671. The latter example may be Hemasadh [671.23- 
24]. The Hemasadhana indeed prescribes a number of rites to find hidden treasure: 
[672.20-22], [674.6-8], [677.25-29], [683.1-7], [683.16-18], [683.18-19], [684.8-10], 
[684.17-18], [686.10-11], [689.11-13], [702.28-703.1], [710.2-3], [718.10-12]. 

53 When the performer enters a forest, all goods (dravya) and medicinal herbs stand 
before him (sarvadravyausadhaya svarupena tisthanti). Amoghapasa 22b. 4 and 
28b. 3 describe two rites which bring about almost the same results. In the long 
chapter of Amoghapasa 21b.5-23a.2 there is described a magical practice concern- 
ing the prescription of a forest (vanavidhisadhana), through the performance of 
which the practitioner opens the door of a forest and goes to the world of the nagas. 
According to Hemasadh [695.5-8] entering a cave (bilapravesa) and entering a for- 
est (vanapravesa) are referred to as the results of the recitation of a certain mantra. 
In Rajatarangim 3.465-470 King Ranaditya obtained a mantra called Hatakesvara 
and entered a cave where he enjoyed the love of the daitya women (Stein 1900: 

54 Proper rainfall was of the greatest concern for people from the time of the Veda 
(Rau 1957: 90-92). RV 10.98 is a hymn praying to Brhaspati for rain for the sake of 
King Santanu. RV 7.103 , known as the frog-hymn, is also a hymn to pray for rain 
(Kajihara 2002: 29). AV 4.15, 6.22, and 7.18 also serve the same purpose (Bloom- 
field 1899: 80). For a ritual to pray for rain among the kamyesti, see Caland 1908: 
13, 38, 129-134. PB 6.10.15-18 is a kamyasoma for obtaining rainfall. In the 
Brahmana literature we find sporadic statements that certain ritual actions are 
devices to cause to rain or to prevent rain falling, for example, in KS 26.6 [128.19], 
27.1 [138.13-15], MS 3.9.4 [118.14-16], 4.5.5 [71.6-9], 4.6.9 [92.9-11], TS,,, SB, JB 1.117 [50.14-21], 1.184 [76.30] and so on. These 
are surely only a small portion of such statements. For rites for rain described in 
the Grhyasutras and auxiliary texts, see Gonda 1980: 44, 398-399. We can add 
the following passages: KausS 41.14, AgnGS 2.5.11 [90.17-91,23], AVPS 30b.l.l7, 
65.3.8, Santikalpa 2.17.2 (aindrisanti), Rgvidh 2.155-156. From the Puranas I col- 
lected only two cases: Devibhagavata Purana 11.24.57-58 and Bhavisya Purana 
4.139.41-42ab. The Amoghapasa mentions rites for rain also in 19a. 1 and 28b. 2. 
Only Hemasadh 55 [684.11-12] and 55 [719.8] describe rites for rain. We are left 
with the impression that compared with rites to get dinaras, this text pays very 
little attention to rainfall. 

55 Amoghapasakalparaja 48b. 6-7 [23.6-8] also prescribes a rite for the protection of 
the house. Rgvidhana 4.131cd-135 is a rite for the pacification of a dwelling place 
(vastusamana) (see above, p. 5) and Samavidhana 3.3.6-7 also teaches the same 
rite (see above, p. 10). These rites have been already dealt with in the Grhyasutras: 
see, for instance, KausS 23.1-8, JaimGS 2.6 [31.10-32.1], GobhGS 3.9.4, ManGS 
2.11.7, BodhGS 3.4.1-21, BharGS 2.4-6 [34.8-37.15], HirGS 1.8.12-15, AgnGS 


Genesis and Development of Tantrism 

rain (ativrsti): 56 44b.4-5 [60.8-10], to stop wind, cloud and lightning: 57 44b.5 

3-2-1 absence of fear (abhaya): 

Great magic boundary (mahasimabandha): 58 44b.5 [60.10] 

3-2-3 release from various evil conditions: 

From all sins such as obstruction (papavarana) 59 and from all diseases and 
all enemies: 44b.2-3 [59.23-28], from all vinayakas: 44a.3 [58.10-12] 

5-2 rites for battle: 

Immobilization of an army (sainyastambhana): 44b.5 [60.12], smashing a 
hostile army (paracakranidhapana): 44b.5 [60.12-13] 

7 sorcery (abhicara): 

abhicaruka: 44b. 4 [60.4-5], attraction of a yaksakanyalyaksini: 44a. 4-6 
[58.16-26], attraction of Sakra, Brahma, Visnu, Mahesvara, etc.: 44b.6 
[60.17-19], attraction of the moon and the sun: 44b.6 [60.19-21], subjugation 
of a woman, a man, a boy or a girl: 44a.7 [59.7-9], subjugation of a king, a 
minister, a village, a town, etc.: 44a.2 [58.1-4], subjugation of a king and 
all that belongs to him: 44a.7 [59.4-7], subjugation of a mahdrdja with his 
harem and family: 44b.6-7 [60.21-25], subjugation of the peoples of the four 
castes (caturvana loka): 44a.7-44b.l [59.1-3], suppression (nigraha) of all 
bhutas, grahas, yaksas and raksasas: 44b.6 [60.25-28] 

As in the case of the Vinasikhatantra, the variety of the rites here are also very 
limited and the results of these rites are in many cases exaggerated and of a su- 
pernatural character. It seems as if the tendency from kdmas to siddhis reaches 
its climax in the Amoghapasakalparaja. 

2.4.1-2 [61.4-15], AgnGS 2.5.9 [88.5-19], Santikalpa 22.3-5, BodhGSS 1.18 [230- 
231], HirGSS 1.6.2 [76.16-28]. See Kane 1977: 790-791. 

56 The Amoghapasakalparaja also hands down the same rite in other places, such 
as 19a.l, 19a.2-3, 28b.2 and 29b.5. KausS 38.7 is a rite against heavy rain. 
Arthasastra 9.7.84 refers to a similar rite. 

57 Manjusrimulakalpa 55 [692.2-3] teaches also a similar rite to stop wind and cloud. 

58 The rite for making a magic boundary seems to have been a very well- 
known rite in esoteric Buddhist texts as one of the preparatory acts, see e.g. 
Amoghapasakalparaja 3b.3-4, 3b.5, 6b.6-7, 19a.3, 31a.6, 46a.3-4, 48a.6-7, 48b.2- 
3, 50b.4, 65b.3-4, Manjusrimulakalpa 55 [693.16-18], [691.27-692.1], [695.5-8], 
[710.15-16]. See also Suvarnaprabhasottamasutra, Sarasvatiparivarta 106.5- 
107.3 and Susiddhikara Sutra 18 (Giebel 2001: 201-202). 

59 Similar rites for diminishing obstructions caused by previous acts (karmavarana) 
are laid down in Manjusiimulakalpa 55 [673.8-9] and [694.24-29]. 


From kamas to siddhis 


Arbman, E., 1922, Rudra: Untersuchungen zum altindischen Glauben und Kul- 

tus, Uppsala: A.-B. Akademiska Bokhandeln. 
Bhat, M.S., 1987, Vedic Tantrism: A Study of Rgvidhana of Saunaka with Text 

and Translation, Delhi: Motilal Banarsidass. 
Bloomfield, M., 1899 (1978), The Atharvaveda, Strassburg: Verlag von Karl J. 

Triibner (Reprint, New Delhi: Asian Publication Services, 1978). 
Buhnemann, G., 2000, "Buddhist Deities and Mantras in the Hindu Tantras: II 

The Srividyarnavatantra and the Tantrasara," Indo-Iranian Journal 43, 

pp. 27-48. 
Caland, W., 1908, Altindische Zauberei: Darstellung der altindischen "Wun- 

schopfer", (Reprint, Wiesbaden: Dr. Martin Sandig oHG, 1968). 
Bisschop, P. and A. Griffiths, 2007, "The Practice involving the Ucchusmas 

(Atharvavedaparisista 36)," Studien zur Indologie und Iranistik 24, pp. 

Dumont, P.-E., 1962, "The Animal Sacrifice in the Taittiriya-Brahmana, The 

Part of the Hotar and the Part of the Maitravaruna in the Animal 

Sacrifice, TB 3.6 with Tr.," Proceedings of the American Philological 

Society 106-3, pp. 177-182. 
Dumont, P.-E., 1969, "The kamya Animal Sacrifice in the Taittiriya-Brahmana, 

TB 3.8 with Tr.," Proceedings of the American Philological Society 113-1, 

pp. 34-66. 
Einoo, S., 1986, "The Interpretation of the Caturmasya Sacrifice according to the 

Ancient Indian Brahmana Literature," Bulletin of the National Museum 

of Ethnology 14-2, pp. 1001-1068 (in Japanese). 
Einoo, S., 1988, Die Caturmasya oder die altindischen Tertialopfer: Dargestellt 

nach den Vorschriften der Brahmanas und der Srautasutras, Monu- 

menta Serindica no. 18, Tokyo: Institute for the Study of Languages and 

Cultures of Asia and Africa, Tokyo University of Foreign Studies. 
Einoo, S., 2005, "Mrtyumjaya or Ritual Device to Conquer Death," in L. Gbhler, 

ed., Indische Kultur im Kontext, Rituale, Texte und Ideen aus Indien und 

der Welt: Festschrift filr Klaus Mylius, Wiesbaden: Harrassowitz Verlag, 

pp. 109-119. 
Falk, H., 1986, Bruderschaft und Wiirfelspiel: Untersuchungen zur Entwick- 

ungsgeschichte des Vedischen Opfers, Freiburg: Hedwig Falk. 
Gampert, Wilhelm, 1939, Die Siihnezeremonien in der altindischen Rechtsliter- 

atur, Prag: Orientalisches Institut. 
Giebel, R. W, 2001, Two Esoteric Sutras: The Adamantine Pinnacle Sutra, The 


Genesis and Development of Tantrism 

Susiddhikara Sutra, Berkley: Numata Center for Buddhist Translation 
and Research. 

Gonda, J., 1951, The Rgvidhana: English Translation with an Introduction and 
Notes, Utrecht: N.V. A. Oosthoek's Uitgevers Mij. 

Gonda, J., 1980, Vedic Ritual: The Non-solemn Rites, Leiden-Koln: E.J. Brill. 

Gonda, J., 1989, Prayer and Blessing: Ancient Indian Ritual Terminology, Lei- 
den / New York / K0benhavn / Koln: E.J. Brill. 

Goudriaan, Teun, 1973, "Tumburu and his Sisters," Wienerzeitschrift fur die 
Kunde Siidasiens 17, pp. 49-95. 

Goudriaan, Teun, 1978, Maya Divine and Human, Delhi: Motilal Banarsidass. 

Goudriaan, Teun, 1985, The Vinasikhatantra: A Saiva Tantra of the Left Cur- 
rent, Edited with an Introduction and a Translation by T Goudriaan, 
Delhi: Motilal Banarsidass. 

Kajihara, M. 2002, The brahmacarin in the Veda: The Evolution of the "Vedic 
Student" and the Dynamics of Texts, Rituals, and Society, A thesis pre- 
sented to the Department of Sanskrit and Indian Studies, Harvard Uni- 

Kane, P. V., 1977, History of Dharmasastra, Vol. V, Part II, Poona: Bhandarkar 
Oriental Research Institute. 

Kimura, Hideaki, 2005, "Homa Rituals in the ahutividhi Chapter of the 
Amoghapasakalparaja," Journal of Indian and Buddhist Studies 53-1, 
pp. (38)-(44). 

Knauer, F. 1884, Das Gobhilagrhyasutra, herausgegeben und ubersetzt, Leipzig: 
Commissionsverlag von Simmel & Co. 

Kulke, Hermann, 1991, "Gramakama: "das Verlangen nach einem Dorf", 
Uberlegungen zum Beginn fruhstatlicher Entwicklung im vedischen 
Indien," Saeculum 42, pp. 111-128. 

Magoun, H. W., 1899, Atharvaveda Parisista 35, Edited with translation and 
commentary by H. W. Magoun, American Journal of Philology. X, pp. 

Mylius, K., 2000, Das altindische Opfer: Ausgewahlte Aufsatze und Rezensionen, 
Wichtrach: Institut fur Indologie. 

Oertel, H., 1994, Kleine Schriften, herausgegeben von Heinrich Hettrich und 
Thomas Oberlies, Stuttgart: Franz Steiner Verlag. 

Oldenberg, H., 1892 (1967), The Grihya-Sutras: Rules of Vedic Domestic Cere- 
monies, Part II = The Sacred Books of the East, Vol. XXX, Oxford: Oxford 
University Press (Reprint, Delhi: Motilal Banarsidass, 1967). 

Sanderson, A., 2001, "History through Textual Criticism in the Study of 
Saivism, the Pancaratra and the Buddhist Yoginitantras," in F Grimal, 


From kamas to siddhis 

ed., Les sources et le temps: a colloquium, Pondicherry 11-13 January 

1997, Pondichery: Institute Francais de Pondichery, pp. 1-47. 
Stein, M.A., 1900, Kalhana's Rajatarangini: A Chronicle of the Kings of Kasmir 

(Reprint, Delhi: Motilal Banarsidass, 1979). 
Thite, G., 1978, "Vijhana: A Kind of Divination-rites in the Vedic Literature," 

Wienerzeitschrift fiir die Kunde Sudasiens 22, pp. 5-17. 
Tbrzsok, J. 2000, "Tantric Goddesses and their Supernatural Powers in the 

Trika of Kashmir (bhedatraya in the Siddhayogezvarimata)," Rivista 

degli Studi Orientali, LXXVIII, pp. 131-147. 

Institute of Oriental Culture, University of Tokyo 


The Saiva Age 
— The Rise and Dominance of Saivism During the Early Medieval Period — 

Alexis Sanderson 

The early medieval period, from about the fifth century to the thirteenth, 
saw a decline in the role of Srauta sacrifice in the religious ceremonies under- 
taken by Indian rulers. But it was not that kings turned aside from the brahman- 
ical tradition in a fundamental sense. They continued to uphold the brahmanical 
social order of the castes and disciplines (varnasramadharmah) and they were 
commonly commended in inscriptions from the fifth to the eighth centuries for 
having rigorously imposed it on their subjects. We see this in the case of the 
Maukhari Harivarman in the fifth century, the Maharajadhiraja Gopacandra of 
Vahga and the Parivrajaka Maharaja Samksobha of Dabhalarajya in the sixth, 
the Pusyabhuti Prabhakaravardhana of Kanyakubja, Bhaskaravarman of Prag- 
jyotisa, the Maitraka Kharagraha II Dharmaditya of Valabhi, the Gurjara Dadda 
III of Bharukaccha in the seventh, and the Licchavi Sivadeva of Nepal at the 
turn of the seventh and eighth. 1 The same claim is seen in the account of the 

1 CII 3, p. 220, 11. 1-2: varndsramavyavasthdpanapravrttacakrah '[Harivarman], 
who set in motion the establishing of the distinctions between the caste-classes 
and disciplines'; Rajaguru 1962, 11. 6—9: varndsramavyavasthdhetuh *sdksdd 
(corr. Rajaguru -.saksad Ep.) dharma Hvopdttajanmd (corr. : ivopdntajanmd Ra- 
jaguru) ■ • .paramamdhesvaro mahdrdjddhirdjasrigopacandra- 'Maharajadhiraja 
Gopacandra, entirely devoted to Siva, who caused the distinctions between the 
caste-classes and disciplines to be established, as though he were Dharma incar- 
nate'; EI 8:28, 11. 11—12: varndsramadharmasthdpandbhiratena (Samksobha); 
EI 4:29, 1. 3: varndsramavyavasthdpanapravrttacakrah (Prabhakaravardhana); EI 
12:13, 11.34-35: bhagavatd kamalasambhavendvakirnavarndsramadharmapravi- 
bhdgdya nirmito bhuvanapatir 'King [Bhaskaravarman], created by Brahma him- 
self to separate the caste-classes and disciplines that had abandoned their du- 
ties'; CII 3, pp. 173ff., 11.43-44: saksad dharma iva samyagvyavasthapitava- 
rndsramdcdrah '[Kharagraha II Dharmaditya], who established the observances 
of the the caste-classes and disciplines, as though he were Dharma in visi- 
ble form'; CII 4i:21, 11. 7—9: mahdmunimanupranitapravacanddhigamavivekasva- 
dharmdnusthdna*pravino (em. MlRASHl :pravani Ep.) varndsramavyavasthon- 
mulitasakalakalikdldvalepa<h> '[Dadda III], who uprooted all the taints of this 
[degenerate] age of Kali by establishing the separation of the caste-classes and dis- 
ciplines, well-versed in the execution of his duty [as the king] through discriminat- 
ing understanding of the teachings authored by the great sage Manu'; LKA 140, 
11. 1-2: suvihitavarndsramasthitir licchavikulaketur . . . mahdrdjddhirdjasrlsivade- 
vah 'Maharajadhiraja Sivadeva, war-banner of the Licchavi dynasty, who correctly 
established the system of the caste-classes and disciplines'; LKA 143, 1. 1: sam- 
yagviracitasakalavarndsramavyavasthah '[Sivadeva], who correctly fashioned the 
system of the distinct castes and disciplines'. 


Genesis and Development of Tantrism 

history of Kashmir before the advent of the Karkota dynasty in the seventh cen- 
tury given in the twelfth by the Kashmirian historian Kalhana. His chronology 
for this early phase of his country's history is confused, but it is likely that we 
should assign to the fifth or sixth century the king Gopaditya whom he com- 
mends for having restored the first and perfect Age through his regard for the 
castes and brahmanical disciplines. 2 He also reports a popular belief of his time 
that in order to promote the orthodox brahmanical social order the Hephthalite 
Mihirakula, who ruled Kashmir in the early sixth century, had settled natives of 
Aryadesa in his kingdom, which was then, we are told, devoid of the true religion 
(dharmah), being overrun by Dards and Tibetans. 3 

Seeing these claims of the royal imposition of the varnasramadharmah one 
thinks of the non-geographical definition of territory fit for brahmanical rites 
(yajhiyo desah) formulated by Manu's commentator Medhatithi during the ninth 
or tenth century, namely that it is any land in which a conquering brahmanical 
king settles the four caste-classes and imposes on the rest of the population the 
status of untouchables (candalah). This definition served, I propose, to accommo- 
date the fact of the territorial expansion of brahmanical society into new regions 
that was one of salient features of the early medieval period. 4 

2 Rdjatarahginl 1.339: jugopa gopaditya 'tha ksmdm sadvipdm taddtmajah | varnd- 
sramapratyaveksddarsitddiyugodayah 'Next his son Gopaditya protected the earth 
and its continents, causing men to experience the arising of a [newl First Age 
through his attention to [the maintenance of] the caste-classes and disciplines'. 

3 Rdjatarahginl 1.312— 313b: dkrdnte ddradair bhauttair mlecchair asucikarmabhih 
| vinastadharme dese 'smin *punydcdrapravartane (conj. :pravartanam Ed.) | drya- 
desydn sa samsthapya vyatanod ddrunam tapah '*In order to (conj.) promote pious 
observance in this land that had been overrun by barbarians of impure conduct, 
Dards and Tibetans, and [sol had lost the [brahmanical] Dharma, he settled [brah- 
mins] of Aryadesa. Thereafter he performed a terrible penance'. Stein (1979, p. 46), 
no doubt faithfully reproducing the reading of the codex archetypus, gives punyd- 
cdrapravartanam rather than punyacdrapravartane and this leaves him no alter- 
native other than to take not only ddrunam tapah but also this as the object of the 
verb: "he performed a terrible penance, and re-established pious observances". But 
the reading is unacceptable. For even if one can believe, as I cannot, that punydcdra- 
pravartanam vyatanot is not too inelegant an expression for an author of Kalhana's 
calibre, there remains the fact that it requires us to believe also that vyatanot gov- 
erns two objects even though the conjunction necessary for this interpretation is 
lacking. I have therefore emended to punyacdrapravartane, which, taken as an 
instance of the use of the locative of purpose (nimittasaptami), yields an entirely 
appropriate meaning and supposes a scribal error that is readily explained by the 
ease with which readers of the Kashmirian script can mistake -e for -am, the com- 
mon substitute for -am. Furthermore, Stein's rendering of dryadesydn samsthapya 
as "after killing the inhabitants of Aryadesa" is, in my view, much less probable 
than the alternative adopted here, which is to take the verb form samsthapya in its 
contextually more appropriate meaning, namely 'having settled'. 

4 See Sanderson 2005a, pp. 400-401, citing Medhatithi, Manusmrtibhdsya p. 80, 


The Saiva Age 

Thus the first centuries of this period are presented in our sources as 
marked not by the decline of brahmanism but rather by its imposition, rein- 
forcement, and expansion. Moreover, there is abundant epigraphical evidence 
of kings throughout this time bringing Vaidika brahmins into their kingdoms 
by making them grants of tax-exempt land, 5 thereby extending the penetration 
of brahmanical culture while facilitating the administration of their territories 
and promoting agricultural development. 6 

Nonetheless, while kings continued to accept their role as the guardians of 
the brahmanical order (varnasramaguruh), their personal religious commitment 
generally took the form of Buddhism, Jainism, or, more commonly, devotion to 
Siva, Visnu, the Sun-God (Surya/Aditya), or the Goddess (Bhagavati), the deities 
of the new initiatory religions, allegiances that were commonly declared in their 
inscriptions by the inclusion amid their royal titles of epithets that mean 'entirely 

11.24-26 on 2.23: yadi kathamcid brahmdvartddidesam api mlecchd dkrameyuh 
tatraiva <ca> <svadharma?>vyavasthdnam kuryuh bhaved evdsau mlecchadesah. 
tathd yadi kas cit ksatriyddijdtlyo rdjd sddhvdcarano mlecchdn pardjayec cdtur- 
varnyam vdsayen mlecchdms cdrydvarta iva cdnddldn vyavasthdpayet so 'pi sydd 
yajhiyah 'If somehow foreigners were to invade such [pure] regions as that between 
the Sarasvati and Drsadvati rivers (Brahmavarta) <and> impose <their religion?>, 
then even they would definitely become foreign lands [unfit for sacrifice]. By the 
same standard, if some king belonging to the Ksatriya or other [suitable caste- 
class] and of orthodox [brahmanical] observance were to conquer foreigners [in their 
lands], settle communities of the four caste-classes [there], and impose on those for- 
eigners the status of untouchables, just as in the brahmanical heartland of India 
north of the Vindhyas (Aryavarta), then those territories too would be fit for the 
performance of [Vaidika] sacrifices'. 

' On the duty of the king to donate [tax-free] land and other valuables to learned 
Vaidika brahmins (viprdh, srotriydh) see, e.g., Ydjnavalkyasmrti 1. 315-320; 1. 323: 
ndtah parataro dharmo nrpdndm yad randrjitam | viprebhyo diyate dravyam . . . 
'There is no higher religious obligation for kings than that of bestowing the wealth 
they acquire through war on learned Vaidika brahmins ...'; Visnusmrti 3.81-82: 
brdhmanebhyas ca bhuvam pratipddayet . . . 'He should bestow land on brah- 
mins'. On the king's duty not to tax learned Vaidikas see Manusmrti 7.133ab: 
mriyamdno 'py ddadita na rdjd srotriydt karam 'Even though dying [through 
poverty] a king may not levy a tax from a learned Vaidika'. The giving of land 
to learned brahmins is already advocated at length as the king's religious duty in 
the Mahdbhdrata {Anusdsanaparvan, Adhydya 61); and that passage includes an 
injunction that it should be read to the king immediately after his consecration 
(13.61.36: abhisicyaiva nrpatim srdvayed imam dgamam). 

' For a study of land-grants to brahmins (brahmadeyam, agrahdrah, sdsanam) dur- 
ing our period in a particular region, Orissa and northern Andhra Pradesh, see 
Singh 1994, pp. 123-243. For the same in the Far South in Pallava and Cola times 
see Karashima 1984, especially pp. 3, 36-40, and 129; and Stein 1994, especially 
pp. 63-89 and 141-172. The migration of groups of north-Indian Vaidika brahmins 
as recipients of royal grants is the subject of Datta 1989. See also Dutta 1995, 
pp. 97-118 on the practice and implications of land-grants to brahmins in northern 
India c. 400-700. 


Genesis and Development of Tantrism 
devoted' to the founder or deity of whichever of these religions they favoured. 


Among these alternatives devotion to Siva was the most commonly adopted. 
During this period the epithet paramamahesvarah 'entirely devoted to Siva' is 
the most frequently encountered in declarations of the religious adherence of 
rulers; 7 and of the many temples surviving or reported in inscriptions that were 
established by rulers and other notables from the late sixth century onwards in 
the subcontinent, the Khmer realm, the Cham kingdoms of Indo-China, and the 
kingdoms of Java and Bali, those dedicated to the worship of this god are much 
the most numerous. 8 

The preponderance of Saivism during this period is also revealed by evidence 
that all the other religious traditions competing for patronage were colonized or 

7 The royal epithet paramamahesvarah first appears in the epigraphical record in 
the fourth century in Andhra, in an inscription of the Salankayana Maharaja De- 
vavarman of Vengipura (EI 9:7, 11. 1-7), probably the earliest of the Salarikayanas 
in our records since this inscription alone is in Prakrit: sirivijayavehgipura bha- 
gavato cittarathasamipadanujjhatassa bappabhattarakapadabhattasya parama- 
mahessarassa salahkayanassa asamedhaydjino maharajasirivijayadevavammassa 
vayanena . . . 'From victorious Vengipura: by the command of the Salankayana, 
who has performed the Asvamedha sacrifice, the venerable Maharaja Vijayadeva- 
varman, favoured by [his kuladevata, the Siva] Citrarathasvamin, loyal to [his] 
venerable father, entirely devoted to Siva . . . '. It is mostly found in inscriptions but 
occasionally appears on coinage. Thus the coins of Krsnaraja, the Kalacuri king 
of Mahismati, who ruled c. 550-575, have on their reverse, (with corrected ortho- 
graphy): paramamahesvara matapitrpadanudhyata srikrsnaraja (MlRASHl, CII 4i 
p. clxxxi). This is the standard term, as is confirmed by its use in literary sources. 
But we also find the synonym atyantamahesvarah (e.g. CII 5:3, 1. 8: Vakataka 
Prthivlsena I, late fourth century), and, though very rarely and not to my knowledge 
in any inscription, paramasaivah (Petech 1984, pp. 57 and 61: the twelfth-century 
Nepalese kings Indradeva and Anandadeva in the colophons of manuscripts). That 
the Taddhita mahesvarah is to be understood as formed from the name Mahesvara 
in the meaning 'devoted to Mahesvara' (mahesvarabhaktah), i.e. 'devoted to Siva', 
is proved beyond doubt by the occurrence in inscriptions of analytic renderings of 
parallel terms. Thus where the affiliation is with Visnu (/Bhagavat) we see not 
only paramabhagavatah but also param bhagavadbhaktah and in the case of the 
Sun-god (Surya/Aditya) we see both paramasaurah and paramadityabhaktah. And 
there are some cases in which the name of the deity precludes any but the analytic 
form. Thus where the deity is the Goddess or Mahabhairava we see param bha- 
gavatibhaktah and atyantasvamimahabhairavabhaktah. For all these epithets see 
Mirashi CII 3, pp. 253-254, n. 3. 

8 This can readily be observed by perusing the published volumes of EITA. On the 
pre-eminence of Saivism among the Khmers up to the fall of Angkor see SANDER- 
SON 2005a, pp. 402-421. For the situation in Karnataka, where Saiva foundations 
greatly outnumbered others throughout the perod from the fifth to fourteenth cen- 
turies see p. 298. For Kashmir see p. 298, and for Andhra see p. 300. 


The Saiva Age 

profoundly influenced by it. In the first part of this study I shall present this 
evidence for each religion in turn, but with particular attention to Buddhism. In 
the second I shall attempt to explain the factors that enabled Saivism to attain 
this dominant position. 

The Incorporation of Saktism 

The worship of the Goddess was progressively subsumed within Saivism, 
being promoted by its adherents as a higher form of that religion. 9 The Saiva 
mainstream was, as one might expect, focused on Siva. This is so in the 
earliest forms of the religion, which later Saivas would call the Atimarga, 
practised by such Saiva ascetics as the Pancarthikas, Lakulas, and So- 
masiddhantins, and it continued to be so in the Siddhanta, the core tradition of 
the Mantramarga that emerged out of the Atimarga from about the fifth century 
onwards, first in the corpus of Nisvasa scriptures 10 and then in a number of 
others, notable among which are the Paramesvara (Pauskaraparamesvara), 
the Svayambhuvasutrasamgraha, the Rauravasutrasamgraha, the Matanga- 
paramesvara, the Sarvajnanottara, the Kalottara in a number of redactions, the 
Kirana, the Parakhya, the Mrgendra, the Brhatkalottara, the Mayasamgraha, 
the Devydmata, and the Mohacudottara, the last three representing a sub-corpus 
of texts of more restricted application concerned with the rituals of the installa- 
tion of images and the consecration of temples, an area in which officiants of the 
Siddhanta were the dominant operatives. But as this Saiddhantika core grew 
it was progressively surrounded by a diverse array of related liturgical systems 
for the propitiation of various forms of the ferocious deity Bhairava, seen by 
his devotees as a higher, more esoteric manifestation of Siva, and of forms of 
the Goddess seen as embodiments of Siva's divine power (saktih). The Saiva 
scriptures devoted to the cult of Bhairava came to be known collectively as the 
Mantrapitha or Mantra Corpus, headed by the Svacchandatantra, which teaches 
the cult of Svacchandabhairava and his consort Aghoresvari, and the earlier 
among those devoted to cults of Goddesses as the Vidyapitha or Vidya Corpus, 11 

9 On the Sakta elements in Saivism see Sanderson 1988, 1995a, and 2007a. 

10 On the transitional character of the Nisvasa between the Lakula Atimarga and 
the mature Siddhanta see Sanderson 2006, and 2001, pp. 29-31, fn. 32. On the 
probable date of its earliest part see Goodall and Isaacson 2007. 

11 For the use of the term pitham in this context in the meaning 'corpus' or 'collec- 
tion' see Tantraloka 37.18c-19cl, quoting or paraphrasing the lost Anandasastra: 
srimadanandasastradau proktam bhagavata kila || samuhah pitham etac ca dvidha 
daksinavamatah | mantro vidyeti 'The Lord has taught in such scriptures as the 
Ananda that pitham [here means] the corpus [of the non-Saiddhantika Saiva scrip- 
tures!. It is divided into two, to the right and left [respectively], namely the 


Genesis and Development of Tantrism 

headed by (1) the Jayadrathayamala, also known as the Sirascheda, consisting 
of four parts called hexads (satkam) because each is approximately six thousand 
verses in length, which teaches the cult of KalasamkarsanI or Kali in the first 
and those of numerous goddesses worshipped as her esoteric embodiments in 
the remaining three parts, evidently added at a later date — closely related to 
parts of this huge corpus are the scriptures of the Kalikula, Kalikulakrama- 
sadbhdva, Kalikulapancasataka and others, that were the scriptural basis of 
the Kalikula Kali cult known as the Krama, Mahanaya, or Mahartha — , (2) the 
Siddhayogesvarimata, which teaches the cult of the goddesses Para, Parapara, 
and Apara, to which the MalinZvijayottara is related, the scripture taken as the 
foundation of the Trika variant of Sakta Saivism expounded in the Tantraloka 
of the great Kashmirian Saiva Abhinavagupta (fl. c. 975-1025), (3) the Picumata 
or Brahmaydmala, which teaches the cult of the goddess Canda Kapalini and 
numerous related Kalpas, and (4) the texts of the vdmasrotah, of which only the 
Vlnasikha has come down to us intact, which teach the cult of the four goddesses 
Jaya, Vijaya, Jayanti/Ajita, and Aparajita, the sisters of the god Tumburu, 
venerated as an aspect of Siva. 12 

Mantrafpitha] and the Vidyapitha'. The terms 'right' and 'left' assigned to the 
two Pithas follow the common notion that these are the relative positions of the 
male/masculine and female/feminine, Mantras being masculine and the deities they 
embody male and Vidyas being feminine and their deities female. 
12 The distinction in terms of left and right between the two Pithas in the passage 
of the Ananda cited in the preceding footnote must not be confused with that 
between the right current (daksinasrotah) and the left current (vdmasrotah) of 
the Saiva scriptures, which derives from the fact that these are thought to have 
emerged from the right and left faces of the five-faced composite Sadasiva, those 
of Aghora (Bhairava) and the feminine Vamadeva respectively. For of the texts of 
the two Pithas only those of the cult of the four sisters are assigned to the lat- 
ter. The Siddhayogesvarimata and the Picumata are both assigned to the former, 
while according to itself the first Satka of the Jayadrathayamala is a hybrid of 
both (ubhaydtmakam); see SANDERSON 2002, pp. 1-2. Of the other three faces the 
front and rear, the faces of Tatpurusa and Sadyojata, are seen as the source of the 
Garudatantras and Bhutatantras, texts concerned respectively with procedures for 
the curing of the effects of poisons and demonic possession, while the upper face, 
that of Isana, is seen as the source of the scriptures of the Siddhanta, revealing that 
this, unlike the distinction between the two Pithas, is a Siddhanta-centric system 
of classification. It is adapted by the non-Saiddhantika Abhinavagupta as the basis 
of his esoteric account of the nature of the Saiva canon in the Malinivijayavartika 
but only by adding a sixth, upper-upper current (urdhvordhvasrotah) above the 
Siddhanta as the source of the non-dualistic Kaula (Sakta) revelation that he takes 
to be the ultimate ground of the entire canon. Malinivijayavartika 1.160-163b: 
prakrtam brumahe devivisrstds citrasamvidah | ydvat tdvat tad urdhvordhvam 
sroto yad bhedavarjitam || 161 saurabhargasikhddini tatah sastrani tenire | uktam 
bhargasikhayam ca devena paramesthina || 162 urdhvasrotodbhavam jhanam 
idam tat paramam priye | paramadhvaninordhvotthasamvidrupabhidhayind \\ 


The Saiva Age 

To these we may add the scriptures of two later Sakta cults, those of the 
goddesses Kubjika and Tripurasundari. The scriptures of the former, the Ku- 
bjikdmata and related texts such as the Satsahasra, do not claim to be part of 
the Vidyapitha. But they are closely related to, and draw heavily on, the sub- 
corpus of texts within the Vidyapitha that is headed by the Siddhayogesvarimata 
and is asssociated with the Sakta system that would be developed under the 
name of the Trika: the Siddhayogesvarimata itself, the [Trika]kularatnamala, 
the Tantrasadbhava, the Devyayamala, and the Trisirobhairava. Also allied 
in character is the Nityasodasikarnava or Vamakesvarimata, the fundamental 
scripture of the cult of the goddess Tripurasundari. This, which became the 
most widely established of India's Sakta cults, has no direct antecedents in the 
Vidyapitha literature, but is rather an independent development out of an ear- 
lier Sakta tradition of the propitiation of goddesses known as the Nityas in which 
rites for success in love predominated. 13 This early cult was eclipsed by its 

isanavaktraniryatat siddhantad bhedam adisat T shall return now to the matter in 
hand. The nondualistic upper-upper stream is present when the various modes of 
consciousness are [still] in the state of [primal] emission within the Goddess [Para]. 
From this [state of fusion] are created the Saurabhargasikha and other such [nond- 
ualistic (Kaula) scriptures]. And the Supreme Lord has spoken [to this effect] in the 
Bhargasikhd [itself], saying, "This knowledge, O beloved, is the supreme product of 
the upper face". By using the word supreme [here] in reference to the nature of the 
consciousness that has arisen from this upper [face] he shows that he means some- 
thing different from [and superior to] the Siddhanta, which has come forth from the 
face of Isana'. 
1 The distinctness of this tradition is expressed in the Kumdrikhanda of the 
Manthdnabhairava in an account of the hierarchy of the various soteriolo- 
gies. It places those who follow the scripture(s) of the Nityas above those of 
the Atimargic traditions (Mausula, Vaimala, Lakula) and below those of the 
Bhairava corpus comprising the scriptures of the left and right currents. Above 
this it places six Sakta Tantras (parasatkam): three of the Trika (Sadardha 
[=Mdlinivijayottara], Bhairavafkula], and Viravali, then the Kdlikula [texts] of 
the Krama, and finally itself, in two scriptural levels. It is significant that 
it does not put the Nitya cult on the level of its Sakta Tantras or even on 
that of the Bhairavatantras below them; see f. 213r3-7 (Muktisamgrahasutra, 
vv. 108-114c): *musuldyudhahastdndm (em. : mausuldyudhahastdndm Cod.) 
mdydtattvam param padam | suddhajhdnamayd vidyd vaimaldndm param padam 
|| 109 astapramdnavedajhd Idkuldrthavisdraddh | urate pdsupate caiva aisvaram 
paramam padam || 110 navanityagamajnanam sivatattvam param padam 
| tasyordhve *kdrandn (em. : karanah Cod.) panca tyaktvd urdhvam tu 
bhairava<h> \\ 111 * sastatantratantrikdnam (?) nityanandam param padam \ 
samandntakaldtltam vamadaksinasamsthitam || 112 pahktikramena mokso 'sti 
satyam nasty atra samsayah | tasya urdhve parasatkam upary upari samsthitam || 
113 sadardham prathamam bhedam bhairavakhyam dvitlyakam | viravali trtlyam 
tu caturtham kalikakulam || 114 tatas tv adyavataram tu tasya urdhvam anahatam 
| srimatkulalikakhyam 'The final destination of the [Mausula Pasupatas,] those 
who carry a club in their hands, is Mayatattva. That of the Vaimala[pasupata]s 
is Suddhavidya[tattva]. For those who are versed in the Lakulafpasupata] doctrine, 


Genesis and Development of Tantrism 

much more successful successor. But nonetheless evidence of it has survived, 
attesting two forms. One is taught in the Nitydkaula, of which a single, in- 
complete manuscript has come down to us in Nepal. Here the goddess Tripura 
is surrounded by a circle of twelve deities comprising eleven Nitya goddesses 
and Kamadeva, the Indian Cupid. 14 The other has been preserved in the eclec- 
tic Manthanabhairava, whose Siddhakhanda contains detailed manual-like in- 
structions for a Sakta cult of Tripura and nine Nityas with Kamadeva as her 
consort. 15 The earlier prominence of the Nitya cult is indicated by the fact that 
a syncretistic text of the cult of Kubjika, the *Cincinimatasarasamuccaya, con- 
tains a section drawn from the Nitydkaula, or from some lost text closely re- 
lated to it, in which it sets out this cult as the 'teaching of the southern or- 

mastering the eight Pramdna scriptures, and for [those, the Pancarthikapasupatas, 
who engage] in the Pasupata observance, it is [the Tattva] of Isvara. For those 
versed in the scriptural tradition of the Nine Nityas it is Sivatattva. Above that 
is Bhairava, transcending [all] the five Causes: Brahma, Visnu, Rudra, Isvara, and 
Sadasiva]. This, eternal bliss, is the final destination of the Tantrikas of the Tantras 
of the eight [Bhairavas] [v. 132: the Niskala-Svacchandabhairava, the Sakala- 
Svacchandabhairava , the Bahurupabhairava, the Aghorisabhairava , the Vyddhi- 
bhaksabhairava, the Candragarbhabhairava, the Vijndnabhairava, the Tumburu- 
bhairava (perhaps =the Vindsikha), and the Amrtesvarabhairava (=Netratantra)]. 
It is beyond the [universe] that culminates in Samana and is established in [the 
two divisions of the Bhairavatantras, those of] the left [current (vdmasrotah)] and 
[those of] the right [daksinasrotah]. The truth — there is no [room for] doubt in this 
matter — is that liberation is [attained in each these systems but] in the manner 
of ascending a ladder. Above that are the six ascending [divisions of the scrip- 
tures] of Para. The first division is the Sadardha (=Mdlinivijaya, vv. 125a and 
133cd), the second the Bhairava[kula] (=Klinndnvayayoga, v. 134a), the third the 
Virdvali (=Virdvalikuldmndya, v. 134c), and the fourth the Kdlikula [scriptures] 
(=Kdlikdkrama, v. 134d). Above this is the Adydvatdra [of the Pascimamnaya], and 
above that the Anahata [revelation] called Kuldlikd[mndya]' . It is striking that this 
passage omits the Saiddhantikas. It is therefore likely that the text has lost a line 
or verse here. This suspicion is strengthened by the verses that follow. For in these 
the order of systems is repeated with saivam, i.e. the Siddhanta's scriptures, be- 
tween the pdsupatam and the eight Bhairavatantras (v. 128bcd: tat hd pdsupatam 
mahat \ saivam tasya visesam tu bhairavdstakanirnayam). Since the passage also 
omits Sadasivatattva it is probable that it was this level that was assigned to the 
Saiddhantika system in the lost line or verse. To assign the Saiddhantikas to 
Sadasivatattva would, of course, be to disdain their claim that their param padam 
is in fact Sivatattva. 

14 The eleven Nityas of this text are Hrllekha, Kledini, Nanda, Ksobham, 
Madanatura, Niranjana, Ragavati, Madanavati, Khekala, Dravani and Vegavati; 
see Nitydkaula, f. 2r7-2vl. 

15 Manthanabhairava, Siddhakhanda, ff. 186v-231rl. The nine Nityas are 
Kulavidya, Vajresvari, Tvarita, Kurukulla, Lalita, Bherunda, Nilapataka, Mangala 
and Vyomavyapim. The section on Tripura continues to f. 252v and includes the 
text of the Nitydsodasikdrnava. The folio numbers are those of a palm-leaf manu- 
script in private hands, to which I have had access through digital images kindly 
provided by my former pupil and present colleague Dr. Somdev Vasudeva. 


The Saiva Age 

der' (daksinagharamnayah), grouping it with the cult of Kubjika, the cult of 
Kali (Kalikula) in a form attested in the Jayadrathayamala and the related cor- 
pus of the scriptures of the Krama or Mahanaya, and a form of Sakta worship 
agreeing closely with that found in the Trika, calling these the teachings of the 
western, northern, and eastern orders respectively (Pascimagharamnaya, Ut- 
taragharamnaya, and Purvagharamnaya). 

The Saktism of this tetradic schema of the directional Amnayas can be dis- 
tinguished broadly from the earlier Saktism of the Vidyapitha by a marked ten- 
dency to expurgate one of the most conspicuous features of the latter, namely 
its embeddedness in the intensely transgressive tradition of Kapalika asceti- 
cism whose roots lie in the Somasiddhantin division of the Atimarga. Since the 
Saktism of the Amnayas refers to itself as Kaula we may use this term to des- 
ignate these post-Kapalika developments. However, like most terms applied to 
traditions subject to change through time it serves at best to indicate a tendency 
rather than an absolute distinction. For while the cults of Tripurasundari and 
Kubjika adhered to this mode of self-definition and the Trika that developed out 
of the Siddhayogesvarimata also came to do so, 16 the cult of Kali that came to 
constitute the Kaulas' Northern Teaching (uttaramnayah) remained both Kaula 
in its self-definition and firmly Kapalika in its practise. 17 

16 On the anti-Kapalika stance of the mature Trika see Sanderson 2005c, pp. 118- 
119, fn. 74. 

17 For the Kapalika/Mahavratin asceticism of practitioners of the Uttaramnaya, 
that is to say of the Kalikula and Krama/Mahanaya, see Sanderson 2007a, 
pp. 293-294 (Cakrabhanu, Isani, and Jaiyaka), 323 (Cakrapaninatha, author of 
the Bhavopaharastotra). Concerning the date of Cakrapaninatha I was able to 
say in 2007a (p. 417) only that he was earlier than his commentator Ramyadeva, 
who was later than Ksemaraja, which is to say, next to nothing. However, 
since then I have read a Nepalese manuscript, NGMPP CI 14/22, which con- 
tains his Bhavopaharastotra under the title Bhavopaharapuja, and this enables 
us to include him among relatively early authors, since the manuscript is dated 
in 1158/9. To the Kashmirian exponents of the Krama identified as follow- 
ers of the Kapalika observance in 2007a I now propose to add one more. Ac- 
cording to a manuscript of the Chummasamketaprakdsa that I had not seen at 
that time, which contains the final verses of the work that are lacking in the 
one manuscript that I had seen then, the redactor of this text attributed to 
Niskriyananda was one Anantasakti. He is described there as mudradharah (A, 
f. Ilr7— 9): samsarasambhramacayapravibhagabandhasambandhasamksaya*gatir 
(em. : gater Cod.) avikalpamurtih \ saksad anabiladhiya laghuvakkramena 
mudradharas tu vidadhe tad anantasaktih. This expression I take to have 
the same meaning as pancamudradharah 'wearer of the five sect marks [of 
the Kapalika/Mahavratin]'; see, e.g., Svayambhuvasutrasamgraha, Patala 14 
(satsamayabhedah), one of the chapters that is not part of the original work of this 
name, vv. 19-20: caturdasapramanena yuktam kapalam ucyate | kapale ca vratam 
mukhyam sarvapapanikrntanam \ tasmin vratam cared yas tu sanmasan mu- 
ktim apnuyat \ pancamudradharah santah samayacdrapalakah; and Kubjikamata 


Genesis and Development of Tantrism 

In general we may say that these non-Saiddhantika texts with their 
strongly Sakta orientation emerged after the Siddhanta or at least after the 
emergence of its earliest scriptures. Thus, for example, it is clear in my view that 
the Svacchandatantra was redacted after the formation of the Saiddhantika 
Nisvasa corpus, the Tantrasadbhava after the Svacchanda, the Kubjikdmata 
after the Tantrasadbhava, , 18 the first hexad of the Jayadrathaydmala after the 
Kubjikdmata, 19 and the remaining three hexads after the first. 20 However, I see 
no reason to conclude that all that is found in the non-Saiddhantika corpus is 
post-Saiddhantika and some grounds for thinking that some elements may be as 
old or older. This may be the case with the cult of the four sisters of Tumburu. 
For that is known to the Buddhist Dharmakirti (fl. c. 550-650), 21 and the first 
two folios of a post-scriptural text on this cult, the *Devitantrasadbhdvasdra, 
written in learned style in the Arya metre, have survived among the Buddhist 
manuscripts uncovered in Gilgit in 1931. They may be assigned on palaeo- 
graphical grounds to around the middle of the sixth century 22 A second area 

25.31cd: pancamudradharo vapi bhasmanistho digambarah. He is probably 
one with the Anantasakti who wrote the published commentary on the Krama's 
Vatulanathasutra but probably not with the Anantasakti who has left us a com- 
mentary (Visamapadasamketa), as yet unpublished, on the Bahurupagarbhastotra; 
see Sanderson 2007a, p. 344. 

18 See the evidence for this sequence in Sanderson 2001, pp. 20-35. 

19 See Sanderson 2002, p. 1 and note 4 on p. 21. 

20 See Sanderson 2002, p. 2 and note 13 on p. 22. 

21 See Sanderson 2001, pp. 11-13, fn. 10. 

22 No title appears in the surviving fragment of this text. The title assigned here 
is a guess based on the unknown author's description of his work in verses 3 
and 4. There he says that he is extracting the fundamentals (sarah) of the 
Essence of the Tantras (tantrasadbhava!}) of the [four] Goddesses (devinam) that 
had been received from Siva by a sage identified only as the ornament of the lin- 
eage of Atri: 3 atreyavahsatilakenoktam sarvad avapya yat purvam \ suramuni- 
narasuranam devinam tantrasadbhavam || 4 tasmad aham apy adhuna vaksye 
samhrtya saram aryabhih | spastataraksarapahktibhir avisaladhiyam *prabodhaya 
(em. : pravodhata Cod.) 'The Essence of the Tantras of the Goddesses was received 
of old from Siva by the ornament of the lineage of Atri and taught to the gods, sages, 
men, and titans. I in turn have summarized its fundamentals and shall now declare 
them in Arya verses whose lines of syllables will be completely clear in meaning, for 
the instruction of those of modest intellect'. The script is the stage of proto-Sarada 
that Prof. Lore Sander has called Gilgit/Bamiyan type 2 and also Sonderschrift 
1. I stumbled upon the first folio (3221-3222) while searching the facsimiles of the 
Gilgit manuscripts for proto-Tantric Buddhist materials and communicated this un- 
expected discovery to Somdev Vasudeva, then my student, who promptly located 
the second folio (3340-3341) and presented convincing palaeographical arguments 
for the date of the manuscript proposed here (email of 7.12.2000), pointing to the 
presence of the archaic tripartite ya ligature, the occurrence of the old style of hr , 
and the Gupta style ru. The text teaches the Mantras of the four DevTs, who, it says, 
were made manifest at the beginning of creation so that men could attain supernat- 


The Saiva Age 

of the non-Saiddhantika canon that is likely to be very early in origin is that 
of the Yamalatantras assigned to the Vidyapltha, represented in our surviving 
manuscripts by the 12000-verse Picumata, also called the Brahmaydmala. For 
the Skandapurana-Ambikakhanda, whose earliest surviving manuscript was 
completed in 810, lists seven Yamala texts, beginning with the Brahmaydmala, 
as Tantras of the Mother Goddesses (matrtantrani). 23 The date of the text 
itself is still a matter of debate; but it is unlikely to have been composed later 
than the end of the seventh century or earlier than the sixth. 24 It is certainly 

ural accomplishments and liberation (v. lied: prddurbhutd devyah siddhyartham 
muktaye caiva), their ancillaries (ahgamantrdh), their retinue of [four] Dutis and 
[four] Kirikaras (v. 16bc: dutyas sakihkard<h>), Tumburu (v. 17ab: pranavam 
tumburusahitam sarthavaha +), and the Ahkusa (v. 18bc: sapranavam HUM-PHAT- 
viniyuktam ahkusam etat). The Vindsikha, our only complete surviving Tantra of 
the vdmasrotah, teaches the four Devls (vv. 30c-32b), Tumburu (vv. 29c-30b), and 
the Ahkusa (v. 41d etc.), but not the Dutis or Kihkaras. For the fuller pantheon see, 
e.g., Devydmata, f. 40rl:jayd ca vijayd caiva jayanti cdpardjitd | dutibhih kihkaraih 
sdrdham samvrtas *tumburuh (corr : tumburum Cod.) sthitah; Netratantra 11.1— 
27; and Sdraddtilaka 19.87-105b and Tantrasdrasamgraha 23.37-52 (with the four 
Dutis but without the Kirikaras). The expression sdrthavdhah 'the [international] 
trader' in v. 17b (v. 17ab: pranavam tumburusahitam sarthavaha +) no doubt refers 
to Tumburu, who is so described in the Buddhist version of this cult taught in the 
Manjusriyamulakalpa (47.29b, 52a, 54c, p. 413, 1. 12, etc.). According to that source 
the four sisters and Tumburu are to be depicted sailing in a ship with Tumburu at 
the helm (47.24: nauydnasamdrudhd<h> sabhrdtrsahapancamd<h> \ karnadhdro 
*'rthacit (tentative conj. : 'thacit Ed.) tdsdm *tumburundmasamjnitah (em. : turn- 
burur ndma samjnitah Ed.). See also here p. 130. This depiction is also prescribed 
in the Saiva Pihgaldmata, f. 28v5-6 (Citrddhikdra, v. 35): jayddyds cakragds tadvat 
pahktisthd vd likhet \ kramdt ndvdrudhds ca vd likhyds tumburuh karnadhdrakah 
'He should depict Jaya[, Vijaya, Jayanti,] and [Aparajita] forming a circle or in a 
line. Alternatively he may depict them on board a ship with Tumburu as the helms- 
man'. For the early date of this cult see also here p. 129. 

1 See Sanderson 2001, pp. 6-7, fn. 4 and here p. 229 (171.127-130b) and a discus- 
sion of the titles it contains. The oldest manuscript is dated in the year 234. For 
this date and its equivalence to a.d. 810 see Adriaensen, Barker and Isaac- 
son 1994, p. 326. That the era of the date is that of the Licchavi Manadeva 
(=Amsuvarman) was first proposed by Witzel (1986, p. 256, n. 9). The date of 
the commencement of this unnamed era which is seen in Nepalese inscriptions that 
begin during the reign of the Nepalese king Manadeva was determined to fall in 
A.D. 576 on the basis of Tibetan evidence by Luciano Petech (1961). Previously it 
had been assumed that the era was that of Harsa (a.d. 606). 

' Yuko Yokochi has observed (1999a, pp. 81-82) that the icon of the goddess 
MahisasuramardinI seen in texts of the sixth and seventh centuries gives way to 
a new iconic type around the beginning of the eighth century and that this text be- 
longs with the earlier sources in this regard. The same scholar has shown (1999b, 
pp. 68-75) that the description of MahisasuramardinI in 68.10-23 of the text cor- 
responds most closely to the image of MahisasuramardinI from Siddhi-kl-Gupha at 
Deogarh, an example of her Gupta subtype B2. She argues that this was carved in 
the middle of the sixth century or, at the latest, at its the end (pp. 74-75). So, she 
concludes, "the possibility that the text belongs to the same century can no longer 


Genesis and Development of Tantrism 

striking in this regard that it betrays no knowledge of the Siddhanta, its 
Saivism being Atimargic, 25 a circumstance which supports the hypothesis that 
the polarity seen in the Mantramarga between Saivism and Sakta Saivism 
was already present in some form when the former was still in the Atimarga 
stage. 26 Royal devotion to Bhairava certainly goes back before the Siddhanta's 
emergence, being attributed in Vakataka inscriptions to Rudrasena I, who ruled 
c. 335-c. 360, 27 and a copperplate decree issued by Maharaja Bhulunda in 376 
from Bagh (Valkha) in Madhya Pradesh records a grant made to support the 
worship of the Mothers in a temple of those deities established by an officiant of 
the Atimarga, the Pasupatacarya Bhagavat Lokodadhi. 28 

In the light of this evidence that Saktism was extensively incorporated into 
and developed within Saivism it should not be surprising to discover that in spite 
of the prevalence of the worship of the Goddess in early medieval India kings 
identified in inscriptions as devotees of the Goddess (bhagavatibhaktah) rather 
than Siva are very rare. At present I am aware only of Nagabhata, Bhoja, and 
his successor Mahlpala I in the ninth century among the Gurjara-Pratiharas of 
Kanyakubja. 29 

Royal devotion to a goddess, typically as a dynasty's lineage deity (kuladevi, 
vamsadevi, gotradevi), was very common during our period, and such deities are 
often declared in inscriptions to be the source of a king's sovereignty and mar- 
tial might. 30 But this was not sufficient to mark out kings who worshipped such 
goddesses as Saktas. For such worship was common regardless of a king's reli- 

be repudiated" (p. 75). The Gupta type, in one subtype or another, was popular from 
the 5th century to the 8th. 

25 The Skandapurana-Ambikakhanda is not a text of the Atimarga in the sense that 
it was written for initiates in one of its systems. For since it is a Purana its target 
audience is the uninitiated laity. However, the Saivism that it draws on is Pasupata 
rather than Mantramargic. This Atimargic background is conspicuous throughout 
the text; but see particularly Adhyayas 174-183. 

26 Hypothesis first proposed in SANDERSON 1988, p. 667. 

27 See, e.g., the Tirodi plates of Pravarasena II, r. c. 400-c. 450, CII 5:11, 11.3-6: 
atyantasvamimahabhairavabhaktasya . . . maharajasrirudrasenasya. The same for- 
mula appears in all the other surviving copper-plates of this king that are complete 
at this point (CII 5:1, 4, 6-7, 10, 13-14, 18). For these approximate regnal dates of 
Rudrasena 1 1 am following Barker 1997, p. 169. 

28 Ramesh and Tewari 1990:10, 11. 2-6: bhagavallokodadhipasupatacaryapratistha- 
pitakapinchikanakagrdmamatrsthdnadevakulasya pihchikanakam eva gramam 
saha bhadradattavatakagramavatakacchena devagraharamatrnafm] balicaru- 

29 EI 14:13, 11.6, 7, 7-8: param bhagavatibhakto maharajasrinagabhatadevas 
...param bhagavatibhakto maharajasribhojadevas ...param bhagavatibhakto 
maharajasrimahendrapaladevas . . . 

30 For some examples see Sanderson 2007b, pp. 288-290. 


The Saiva Age 

gious affiliation, and it was in any case inconstant, coming to the fore only on cer- 
tain occasions, particularly during the autumnal Navaratra festival that inaugu- 
rates the season of military activity, when they and associated goddesses received 
large-scale animal sacrifices; 31 and when this cult was particularly emphasized 
through the forging of connections with a higher domain of non-periodic, exclu- 
sive devotion, then this domain was that of the esoteric goddesses of the Saiva 
Vidyapitha. 32 

The Etiolation and Subsumption of the Cult of the Sun-God 

As for the cult of the Sun, kings who have been declared in inscriptions 
to be devotees of this god (paramasaurah, paramadityabhaktah, and the 
like) are also few and they are mostly confined to the sixth and seventh cen- 
turies. We have Dharmaraja of Padmakholi in the Ganjam District of Orissa, 
Dharapatta, the Maitraka of Valabhi, Rajyavardhana, Adityavardhana, and 
Prabhakaravardhana, the three successive predecessors of King Harsa of 
Kanyakubja, in the sixth century, and from c. 570 to c. 665 the Gurjara feuda- 

On Navaratra see Sanderson 2005a, p. 371 (fn. 64); 2005b, pp. 255-257; 2007b, 
pp. 263-277 and 294 (fn. 196). For an example of the scale of such annual sacrifices 
see p. 247 below. 

In general we may say that the Saivism of the Mantramarga holds itself aloof from 
the domain of calendrical religion, seeing the recurrent festivals of that domain as 
commemorations of mythic events and therefore as operating on a level of mun- 
dane belief that initiates must transcend. That is the territory of Puranic religion, 
which guarantees various rewards but not the liberation or supernatural effects and 
powers promised to observant initiates into the Mantramarga. Saiva initiates were 
merely required to track the Puranic calendar by intensifying their own regular cult 
on days when uninitiated devotees were celebrating Siva's or the Goddess' activi- 
ties in the domain of myth-based devotion; see, e.g., Tantrdlokaviveka on 28.6d-7b. 
Nonetheless, we see a distinct tendency for the Mantramarga to seep downwards 
into this domain providing Saiva or Sakta Saiva versions of the Puranic rituals 
that mark such major annual festivals as Sivaratri and Navaratra. A Sakta Saiva 
procedure for the celebration of Sivaratri was current in Kashmir, as can be seen 
from the prescriptions set out in the Nityadisamgraha of Rajanaka Taksakavarta 
(ff. 71v-72vl5) from the lost Dutidamara and in the 31st chapter of the Haracar- 
itacintamani of Rajanaka Jayadratha in the thirteenth century, drawing on this 
and the Anantabhaskara. The same can be seen in various regions in the case of the 
Navaratra, also known as the Durgotsava. Among the Newars of the Kathmandu 
valley, the goddess is worshipped in this festival in a Tantric form as Ugracanda 
in Paddhatis that incorporate her among such Mantramargic Sakta deities as 
Siddhilaksmi and Kubjika; see the Newari Navaratrapujavidhi manuscripts A and 
B in the bibliography. For her Tantric worship in this context in the tradition of 
the Paippaladin Atharvavedins of Orissa see Sanderson 2007b, pp. 263-276. In 
Bengal, where Navaratra was and is much emphasized, we see a Smarta procedure 
but one that has been strongly Tantricized in the Durgapujdprayogatattva section 
of the Durgapujatattva of Raghunandana in the 16th century. 


Genesis and Development of Tantrism 

tories of Bharukaccha (Broach). This is explicitly stated in the case of Dadda 
I (r. c. 570-595), and Dadda II (r. c. 620-645); and it is probable in the case of 
Jayabhata II (r. c. 645-665), since it is very likely that the temple of the Sun-god 
Jayaditya at Kotipura near Kavi in the Broach District was founded by him with 
his name (Jaya-). It is also probable in the case of Jayabhata I (r. c. 595-620), 
since this was the religion not only of his predecessor and successor but also 
of his brother Ranagraha. After Jayabhata II the next three kings of this 
dynasty, Dadda III (c. 665-690), Jayabhata III (c. 690-715), and Ahirola (c. 715- 
720), turned to Saivism, declaring themselves paramamahesvarah. In the ninth 
century we have royal devotees of the Sun in Ramabhadra, the immediate prede- 
cessor of the Gurjara-Pratihara Bhojadeva I of Kanyakubja, and Vinayakapala, 
the latter's grandson, and, in the late twelfth and early thirteenth centuries, in 
the Sena kings of Bengal Laksmanasena and Visvarupasena, though the former 
also appears in his inscriptions as a Vaisnava (paramavaisnavah) and, more 
specifically, as a devotee of Narasimha (paramanarasimhah). 33 

It appears that the Sauras, the initiated devotees of the Sun-god, possessed 
their own canon of scriptures, known, like those of the Saivas and the Vaisnava 
followers of the Pancaratra, as Samhitas. A list of eighty-five such texts is given 
in an account of brahmanical, Paficaratrika (Vaisnava), Saura, and Saiva scrip- 
tural authorities, contained in the Saiva scripture Srikanthiyasamhita. No man- 
uscript of this text, which was known to Ksemaraja {fl. c. 1000-1050) and prob- 
ably to Abhinavagupta (fl. c. 975-1025), has come down to us; but I have located 
its long section dealing with the canons of scripture in the Nityadisamgraha of 
Rajanaka Taksakavarta, a Kashmirian digest of scriptural passages bearing on 
the duties of initiated Saivas, compiled at some time after the eleventh century 34 

33 EI 28:16: sahasrarasmipadabhakto (Dharmaraja); EI 31:39B,1. 8: paramaditya- 
bhaktah (Dharapatta); EI 4:29, 11. 1-3: paramadityabhaktah (the predecessors of 
Harsa); CII 4i:16, 1.4: dinakaracaranakamalapranamapanitdsesaduritanivaha- 
(Dadda I); ibid., 1.52: dinakaracaranarcanaratasya (Dadda II); CII 4i:18, 1.9: di- 
nakarakiranabhyarcanaratasya (Ranagraha); CII 4i:21, 1. 13: paramamahesvarah 
(Dadda III); ibid., 11. 16-17: paramamahesvarah (Jayabhata III); CII 4i:24, 11. 20-11: 
paramamahesvarah (Ahirola); EI 5:24,1.5: paramadityabhakto (Ramabhadra); EI 
14:13, 1. 6: paramadityabhakto (Vinayakapala); SlRCAR 1983a:27, 11. 35-38: para- 
masaurah (Laksmanasena); paramasaura (Visvarupasena); EI 12:3,11.23-25: pa- 
ramavaisnava- (Laksmanasena); and SlRCAR 1983a:26,ll. 32— 33: -paramanarasi- 
mha- (Laksmanasena). For the attribution of the temple of Jayaditya at Kotipura 
to Jayabhata II see Mirashi, CII 4i, p. liv. 

34 The list of the Saura Samhitas in the Nityadisamgraha is to be found on ff. 4vll-5r6 
of the codex unicus. A lightly edited transcript of the whole excerpt on the scriptural 
canons has been published as it appears in an apograph contained among the Stein 
manuscripts of Oxford's Bodleian Library by Jiirgen Hanneder (1998, pp. 237- 
268). The verses on the Saura canon are 74-88 in his edition. On the date of the 
compilation of ^the Nityadisamgraha see SANDERSON 2007a, p. 422. 


The Saiva Age 

Unfortunately, no manuscript of any one of these Saura scriptures has surfaced; 
and the decline of Saurism as a distinct tradition, of which this is the conse- 
quence and evidence, is probably to be attributed, at least in part, to a failure to 
continue to attract patronage and so maintain its separate identity as Saivism 
became more influential and encroached upon its territory 

Thus a Saurasamhita of our period sets out the procedure for the worship 
of the Sun and no doubt drew on the Saura tradition. 35 But it assigns itself to 
the canon of the Saiva scripture Vdthula/Kalottara, 36 a text on which it silently 
draws, gives a Saiva account of the place of the Sun in the birth of the uni- 
verse, deriving it through emergence from Siva expressed in a phrase found 
elsewhere in the Saiva scriptures, 37 and insists that Siva and the Sun are in 
essence a single deity 38 Moreover, the worship of the Sun taught in this text 
was included by the Saiddhantika Saivas as a compulsory preliminary (afigam) 
of the regular worship of Siva himself, appearing first in the sources known 
to me in the Siddhdntasdrapaddhati of Maharajadhiraja Bhojadeva of Dhara 
(r. c. 1018-1060) 39 and then soon afterwards, in dependence on that text, in the 

35 A critical edition of this text is being prepared for publication by Dr. Divakar 
Acharya. I am very grateful to him for sending me drafts of this edition. The text 
survives in a Nepalese palm-leaf manuscript with a scribal date that falls in A.D. 
949 (NAK MS 1/1231, NGMPP A1161/6). 

36 Saurasamhita 1.5: nokta purvam tu yd vatsa gopitd saurasamhita | tantre tu 
vathule sd tu rahasyam na prakdsitd. Final colophon: it i vdthule kriydpdde saura- 
samhitdydm .... 

37 Saurasamhita 1.10-12: adrstavigrahdc chdntdc chivdt paramakdranat \ 
kriydsaktir viniskrdntd paratejasamanvitd || 11 dkdse tu yadd hy ulkd 
srstihetor adhomukhi | tasya tejasamdyogdd utpannam tejarupinam || 12 
ddityamanisamyogdd vahnih samjdyate yathd | saktitejasamdyogdd bhdnuh 
sambhavitd tathd. lOab = Pauskarapdramesvara (as quoted by Bhatta Ramakantha 
at Matahgapdramesvaravrtti, Vidydpdda, p. 19, 11.5-6) and Srlkanthlyasamhitd 
(ed. in Hanneder 1998, p. 240, v. 1). 

38 Saurasamhita 1.15: ddityam tu sivam vindydc chivam ddityam eva ca | ndndtvam 
yas tu gaccheta yatnendpi na sidhyati. 

39 Siddhdntasdrapaddhati, MS A, f. 3v5-4v2, MS B, f. 4v6-6r2: OH HRAM HRIM 
SAH iti suryamantrena krtadehasuddhih krtasakalikaranam arghapdtram 
krtvd puspddikam samproksya raktacandanddind surydya mulamantrendrgham 
dattvd siiryam pujayet | tatra ganapatigurupujdnantaram OM AM PRABHUTAYA 
NAMAH iti pithamadhye, OM AM VIMALAYA NAMAH ity dgneyydm, OH AM 
SARAYA NAMAH iti nairrtydm, OM AM ARADHYAYA NAMAH iti vdyavydm, 


punar madhye, OM RAM DIPTAYAI NAMAH purvadale, OM RIM SUKSMAYAI 


NAMAH isane, OM RAM SARVATOMUKHAYAI NAMAH karnikdydm sampujya 
visphurdm mudrdm pradarsya raktavarnavartulatejobimbamadhyastham 
raktavdsasam svetapadmopari sthitam sarvdbharanabhusitam ekavaktram 


Genesis and Development of Tantrism 

Somasambhupaddhati, composed towards the end of the eleventh century. 40 The 
Sambapurana, which teaches the worship of the Sun-god, is also a product, at 
least in its later portions, of a Saiva environment. 41 

Traces of some form of the vanished tradition of the Sauras may have sur- 
vived in the Sakta Saiva literature. For Kashmirian sources know of a Sakta 
cult whose deity was the Sun under the name Vira or Viresvara accompanied 
by the goddess Bhargasikha, citing as its scripture the Kaula Bhargasikha, also 
called Saurabhargasikha, a work for knowledge of whose content we now have 
only a few comments in the Kashmirian literature and a few verses quoted in the 
same, one of which has also been quoted by the east-Indian Buddhist Ramapala 
in his Sekanirdesapahjikd, a fact that demonstrates that this was not a merely 
a local, Kashmirian tradition. 42 The probability that this cult reflects a non- 

dvibhujam svetapahkajapdnim sarvalaksanasampannam samcintya puspair 


ity dvdhanamudrayd samdvdhya sthdpanyd samsthdpya samnidhd<pa>nyd 
samnidhdpya nisthurayd nirodhydrghapddydcamaniydni khasolkina dattvd 
ahgena mulamantrena sdhgam suryam gandhapuspddibhih sampujya padma- 
mudrdm bimbamudram ca pradarsydgneyydm OM AM HRDAYAYA NAMAH, 
(em. :E B:AIH A) jvalinIsikhayai vausat, vayavyam oh hum kavacaya 
HUM, OM BHANUNETRAYA VASAT madhye, purvadicatursu digdalesu OM RAH 
ASTRAYA PHAT ity afigdni sampujya hrdayadinam dhenum netrasya govrsam 
trasanim astrasya ca pradarsya OM SAM SOMAYA NAMAH purvadalagre, OM 


OM KEM KETAVE NAMAH isdnydm iti grahdn sampujya namaskdramudrayd 
prarocya gandhapuspadipadhupanaivedyddi khasolkina dattvd padmamudrdm 
bimbamudram ca pradarsya ksamasvety uccdrya mantrasamuham upasamhrtya 
samhdramudrayd dvddasdntasthitasurydya hrtsthitdya vd niyojayet. ity anena 
vidhind visarjya nirmdlyam arghapdtrodakam ca aisanyam TEJASCANDAYA 
NAMAH I iti suryapujdvidhih. For some detailed evidence of the dependence of the 
Somasambhupaddhati on the Siddhdntasdrapaddhati see SANDERSON 2005a, 
p. 360 (fn. 28). 

40 Somasambhupaddhati, Pt. I, pp. 68-89. 

41 Hazra 1958, pp. 29-108; von Stietencron 1966, pp. 227ff. 

42 See Abhinavagupta, Mdlinislokavdrtika 1.161-162b (160c-161b: ydvat tdvad 
tad urdhvordhvam sroto yad bhedavarjitam || saurabhargasikhadini tatah 
sdstrdni tenire); Tantrdloka 4.255 and 15.280; 32.62: virabhairavasamjneyam 
khecari bodhavardhini \ astadhettham varnitd sribhargdstakasikhdkule; 
Ksemaraja, Sdmbapancdsikdtikd on brahma prathamam atanu in v. 10a: 
prathamam dddv atanu asarlram srlbhargasikhddistanityd akdrapardmarsdtma 
viresvarakhyam brahma brhad brmhakam ca param sdktam dhdma and on 
v. 21: srlbhargasikhdydm api "naisa varno na vd sabdo na caivdyam kaldtmakah 
| kevalah paramdnando viro nityodito ravih \\ ndstam eti na codeti na sdnto na 
vikdravdn | sarvabhutdntaracaro bhanur bharga iti smrta" iti; Svacchandod- 
dyota, vol.4 (Patala 9), p. 55, 11.15-16; and Ramapala, Sekanirdesapahjikd, 
f. 10v2— 3: tad uktam bhargasikhdydm sdkteye tantre na san na cdsat sadasan 


The Saiva Age 

Saiva tradition otherwise lost to us is made somewhat greater by the fact that 
the names Vlresvara and Bhargasikha are applied in Kashmirian sources, both 
Saiva and Smarta, to the Sun-god and his consort at Martandatirtha (modern 
Matan), where King Lalitaditya built his majestic temple of the Sun in the mid- 
eighth century, 43 a site that has been a major pilgrimage site with its own special 
rites for the dead, the Bhargasraddha and Suryabali, down to modern times. 44 
However, it is possible that the application of these names merely reflects the 
pervasive influence of Sakta Saiva esotericism in the wider Kashmirian commu- 
nity in later times. 

There are also strong elements of a solar esotericism in the Kalikula of the 
Jayadrathayamala and the Krama. 45 It is possible that these too may have been 

na tan nobhayojjhitam | durvijheyd hi sdvasthd kim apy etad anuttamam (the 
verse has been silently incorporated by Abhinavagupta as Tantrdloka 2.28 [with 
anuttaram not anuttamam]): Jayaratha identifies this as a quotation from the 
Bhargasikha in his commentary: sribhargasikhdm samvddayati (-viveka, vol. 1, 
Ahnika 2, p. 22). 

43 Rdjatarahgini 4.192; Krishna Deva in EITA, vol.2, part 1, pp. 363-66; plates 
710-721; AIISPL, Accession numbers 20738-20789 and 60003-60051. The 
Mdrtdndamdhdtmya, the praise-text of this site, refers to Surya here as Vlresvara 
(Bhrhgisasamhitd, p. 15: esa viresvaro devah parah paramakdranam; p. 63: 
viresdya namas tubhyam; p. 66: namo virddhiviresa) and makes Bhargasikha the 
first of his Saktis (ibid., p. 12, listing Bhargasikha, Bhlma, Bhasvati and Bhanavi). 
The Sun is also invoked as Vlresvara in the worship of the Grahas that occurs 
among the preliminaries in Saiva rituals in Kashmir; see Kalddiksdpaddhati B 
f. 4v9-10: tadbahir grahdh. tatrddau madhye suryah OM RAM AGNAYE OM HRAM 


Bijas HRAM HRIM SAH are Surya's. His consort is invoked as ViralaksmI here rather 
than as Bhargasikha because in the context of the ritual the pair are superimposed 
on the principal deities Amrtesvarafbhairava] and his consort Amrtalaksmi. 

44 For the Paddhati of these rituals see Karmakanda, vol. 4, pp. 140-205. Here too 
the Sun is invoked as Vira/Viresvara (p. 196): vira viresa devesa namas te 'stu 
tridhatmaka | mahamartanda varada sarvabhayavaraprada .... 

45 See, e.g, Jayadrathayamala 4.4.8-17: sa ravir bhasuradharas tadadhara hi kalika 
| sadare vipuladhara sodasoddyotasannibha || 9 sphuradvamanasangrasaravikl 
srstikarika | sa ravir devatakaro ravir eka<s> tadakrtih || 10 ravih pradipakaloke 
suryamadhyat samutthitah \ raver antargato bhanur bhasayaty akhilam jagat 
|| 11 bhanavi kaulini yd sa tatpunjabharitam jagat \ tatrotpannd mahdmantrd 
bhairavdstdstayonayah || 12 na prakdse na cdkdse nobhaye nobhayojjhite 
sarvdvarananirmukto sarvago bhdti bhdskarah || 13 amrtam prdvrtam yena 
racitam ca kuldkulam | sa ravih suryaturydnte bhrdjate raudraddmarah \\ 14 
svasamvitparamddityanityoditamaricibhih | bhacakram bhdsitam yena sa vai 
kdlanjaro bhavet; CincinTmatasdrasamuccaya, ff. 30v7-21r4 (7.166—172 [Ut- 
taragharamnaya (Kalikula) section]): 166 ravih pradipakaloke suryamadhydd 
vinirgatah \ raver antargato bhanur bhasayaty akhilam jagat || 167 bhanavi kaulini 
yd sa tatpunjabharitam jagat \ tatrotpannd mahdmantrd bhairavdstdstayonayah 
|| 168 ravibhdnumayi devi kaulesi kulagahvari \ ksobhdnandavirdme tu 
pasyate kulasamtatim \\ 169 mahdvyomdrnave saive bhdnavikundamadhyatah 
| tatra pralindh sarve te bhairavdstdstayonayah \\ 170 bhdnavikundamadhye 


Genesis and Development of Tantrism 

constructed on the basis of Saura notions. But it is also possible that they are an 
independent development internal to Saivism. In the absence of properly Saura 
literature it is impossible to be sure. 

The cult of the Sun-god, then, appears to have survived in India after the 
rise of the Saivism only in heavily Saivized Puranic reflexes or subordinated in 
a Saivized form within the Saiddhantika cult of Siva, and, perhaps, in some ele- 
ments within the Sakta Saiva tradition. Only in the Majapahit kingdom of East 
Java do we hear of the survival of adherents of a distinct Saura denomination. 
There a royal charter of c. 1350 tells us that a board of six learned men appointed 
to adjudicate law suits included two adherents of this tradition. 46 

The Decline of Vaisnavism and the Rise of the Tantric Pancaratra 

Following Saiva Models 

Royal preference for Vaisnavism, expressed in inscriptions by the epithets 
atyantabhagavadbhaktah, paramabhagavatah, or paramavaisnavah, all mean- 
ing 'entirely devoted to Visnu', is mostly confined to the period from the fourth 
century to the seventh. The Bhagavata faith was adopted and promoted by 
the Guptas from the first half of the fourth century through to the end of the 
fifth, 47 and it was probably under their influence that it gained a foothold in 
the fifth century among the Saiva Vakataka rulers of Nandivardhana in east- 
ern Vidarbha, through the marriage in the last decade of the fourth century of 
the Vakataka Rudrasena II to Prabhavatigupta, the daughter of the parama- 
bhagavatah Gupta emperor Candragupta II (c. 380-474). 48 Gupta influence may 
also explain the appearance of the Bhagavata faith at the end of the fourth cen- 

tu layacakram svabhdvatah | villne svasvabhdvdkhye tatsvabhdvodayam tatah 
|| 171 bhdvdbhdvadvayottirnd yd virauty asaririni \ sd cidd nihsvabhdvasthd 
surydkuld krsodari || 172 tatsvarupoditam cakram cidbhdnvarkagatisthitam 
pratibimbam ivdbhdti visvagrdsaikalampatam; Kdlikulakramasadbhdva 2.37cd: 
bhdskarair dvidasair yuktd sikhd bhargasya cottamd; Eraka, Kramastotra, quoted 
in Tantrdlokaviveka on 4.165c-167: astoditadvddasabhdnubhdji yasydm gatd 
bhargasikhd sikheva | prasdntadhdmni dyutindsam eti tarn naumy anantdm 
paramdrkakdllm. On the literature of the Kashmirian Kalikula see SANDERSON 
2007a, pp. 250-370. 

46 See here p. 120. 

47 CII 3:8, 11.1-2: paramabhdgavatamahdrdjddhirdjasrikumdragupta-; 11.20-23: 
paramabhdgavato mahdrdjddhirdjasricandraguptas tasya puttras tatpdddnu- 
ddhydto mahddevydm dhruvadevydm utpannah paramabhdgavato mahdrdjddhi- 
rdjasrlkumdraguptas tasya puttras tatpdddnuddhydtah paramabhdgavato mahd- 
rdjddhirdjasriskandaguptah . 

48 On Saivism and Vaisnavism among the Vakatakas of Nandivardhana and the influ- 
ence of the Vaisnava Prabhavatigupta on the religion of this dynasty see Barker 


The Saiva Age 

tury among the Salarikayana kings of Verigipura in Andhra. The earlier kings of 
this dynasty were devotees of Siva in keeping with the norm in this region. But 
Nandivarman II, a younger contemporary of Candragupta II, is styled parama- 
bhdgavatah} 9 Other early Vaisnava kings are the Matharas of Kalihga, 50 the 
Traikutakas of Nasik, Korikana, and Lata, 51 the Sarabhapuriyas of Daksina Ko- 
sala, 52 and the Parivrajaka Maharajas of Dabhalarajya (Dahala) in the fifth and 
sixth centuries, 53 perhaps the early Maukharis of Kanyakubja before the reign of 
Isanavarman (c. 550-76), 54 the Nalas of western Orissa (c. 450+-700), 55 the early 
Calukyas of Vatapi (Badami) in the sixth and early seventh century, 56 and the 
early Pallavas of Kanci up to and including Simhavisnu II (c. 550-610). 57 After 
Pulakesin II and Simhavisnu both the Calukyas and Pallavas were Saivas, 58 as 

49 El 42:11, 11.7—9: bhagavaccitra<rathasvdmya>nuddhydto . . . paramabhdgavatas 
sdlahkdyanavamsaprabhavo vijayavarmmd. For this hypothesis of Gupta influ- 
ence, which rests on slenderer evidence than that of Gupta influence on the 
Vakatakas, see S. Sankaranarayanan in EI 42:11, p. 92. 

50 Tripathy 1997:2: bhagavatsvdmindrdyanapdddnudhydtah; 3: ndrdyanasvdminah 
pddabhaktah paramadaivata<h>. 

51 MlRASHl, CII 4i, p. xliv; CII 4i:8, 11. 1-2: bhagavatpddakarmmakaro . . . mahdrdja- 
dahrasena<h>; CII 4i:9, 11. 1-2, 7-8: bhagavatpddakarmmakarah . . . mahdrdja- 

52 EI 31:35, 11. 1-2; EI 22:6, 11. 3-4; EI 31:18, 1. 3. 

53 EI 8:28. 

54 Of his predecessors Harivarman, Adityavarman, and Isvaravarman, we know that 
the second at least was paramabhdgavatah. 

55 EI 21:24 (Podagadh inscription of the Nala Skandavarman, fifth century) and EI 
26:3 (Rajim stome inscription of the Nala Vilasatuhga, c. 700); Singh 1994, pp. 89- 

56 Kirtivarman I (r. 566-597) completed the Visnu cave-temple at Vatapi. His suc- 
cessor Mahgalisvara-Ranavikranta (r. 597-608) is styled paramabhdgavatah in an 
inscription in the Vaisnava cave 3 at Badami recording the completion of the tem- 
ple, the installation of the Visnu, and the granting of a village (Fleet in Burgess 
1877, p. 363, 11. 5-10; and Fleet 1881 [lithograph]): srimahgalisvararanavikrantah 
. . . paramabhdgavato *layanam (corr. Fleet :layano Ep.) mahdvisnugrham 

. . . krtvd On the Vaisnavism of the early Calukyas before Vikramaditya I (654- 

c. 681) see Bolon 1979, pp. 254-256. 

57 Carudevi, wife of Buddhavarman son of Skandavarman I (c. 330-350) (Maha- 
LINGAM 1988:4, 11. 7-9: gift of land to a temple of Narayana); Simhavarman 
II, c. 436- 477 (MAHALINGAM 1988:8, 11. 15-17: paramabhdgavatah); Yu- 
varaja Visnugopa, mid-fifth century (MAHALINGAM 1988:6: 11. 9-17; MAHALINGAM 
1988:7, 11. 18-21: paramabhdgavatah); Nandivarman I, c. 495-520 (MAHALINGAM 
1988:10, 11.9-10: paramabhdgavatah); Buddhavarman, father of Kumaravisnu 
III (MAHALINGAM 1988:11, 11.6-7: bhagavadbhaktisambhavitasarvakalyanasya); 
Kumaravisnu III c. 520- 540 (MAHALINGAM 1988:11, 11. 12-14: paramabhdgava- 
tah); Simhavarman III c. 540- 550 (MAHALINGAM 1988:12, 11. 14-18: paramabhd- 
gavatah); Simhavisnu c. 550- 610 (MAHALINGAM 1988:76: bhaktydrddhitavisnuh 
simhavisnuh) . 

58 For the Saivism of Calukya Pulakesin II's successors Vikramaditya I (654-c. 681), 
Vinayaditya I (681-696), Vijayaditya (696-733), Vikramaditya II (733-744), and 


Genesis and Development of Tantrism 

were the later Maukharis. 59 

After the seventh century royal Vaisnavism is sporadic, with the prominent 
exception of the Karkotas of Kashmir (c. 625-855/6). The conclusion that this 
dynasty was Vaisnava is not derived from our study of inscriptions, because 
extremely few have survived the centuries of Islamic rule in Kashmir, which 
began in 1339 and ended in 1819. It rests primarily on the testimony of the 
Rajatarangini of the Kashmirian historian Kalhana, who did have access 
to, and did utilize, the local epigraphic record of religious foundations and 
dynastic history 60 From this work we can see that when a king of this dynasty 
established and enshrined a deity, generally with his own name (svanamna), 
it was always a Visnu (-svamin, -kesava), though sometimes images of the 
Sun-god or the Buddha were enshrined in addition. These royal Visnus are 
the Durlabhasvamin (4.6) of Durlabhavardhana (r. c. 626-662), the Tribhu- 
vanasvamin (4.78) of Candrapida (r.c. 712-720/1), the Muktasvamin (4.188) of 
Lalitaditya-Muktapida (725-761/2), his silver Parihasakesava at his new town 
Parihasapura (4.195, 202), his golden Muktakesava (4.196, 201), and a Visnu 
at his new town Darpitapura (4.183), the Vipulakesava (4.484) of Jayapida (r. 
c. 773/4-804/5), and his Caturatmakesava and Anantasayana Visnu at his new 
town Jayapura (4.508), the Amrtakesava established after his death by his 
mother Amrtaprabha to secure the rescue from hell that the sins of his later 
life had made his certain destiny (4.659), and the Visnus established by each of 
the five uncles of Cippatajayapida, who ran the country for thirty-seven years 
during the reign of the puppet king Ajitapida (r.c. 813/4-850/1): Utpalasvamin 

Kirtivarman II (744-c. 753/757) and their construction of the Siva temples at 
Pattadakal and Alampur see EI 32:21, ARE 159 of 1959-60, EI 35:16 and 3:1; and 
the excellent overview in Dagens 1984, vol. 1, pp. 20-24. 

59 On the Saiva affiliation of the Maukharis Isanavarman, Sarvavarman, and Avanti- 
varman see Barker and Isaacson 2004, pp. 32-33; Thaplyal 1985: B 2, 11. 19- 
20; B 3, 11. 7-8; B 5, 11. 7-8. Another lineage that may have been Vaisnava up 
to the early seventh century before turning to Saivism is that of the Varmans of 
Pragjyotisa. Bhutivarman of that line was paramabhagavatah according to his 
Badagariga rock inscription of 553/4 (EI 27:5, 11.1-2): sri paramadaivataparama- 
bhagavatamaharajadhirajasvamedhajdjin[am] srlbhutivarmadevapadanam. But 
his great-great-grandson, Bhaskaravarman (r. c. 600-50), has been described in 
his Dubi copper-plate inscription as having revived Saivism; see SlRCAR 1983a:l, 
11. 109-110): laksmih ksibavilasa[nita]vidhina samskrtya ca svikrtd bhuyo yena ma- 
hesvarasrayanayah sphdyipratdpdrcisd. 

60 Rajatarangini 1.15: drstais ca purvabhubhartrpratistha*vastusasanaih 
(conj. : vastusasanaih Ed.) | prasastipattaih sastrais ca santo 'sesabhramaklamah 
'I have removed all the troublesome errors [of my predecessors] by consulting 
in person the charters that record the [temples and otherl edifices founded and 
consecrated (-pratisthavastu-) by the kings of the past, [their] panegyric donative 
inscriptions, and works of scholarship'. 


The Saiva Age 

(4.695ab), Padmasvamin (4.695cd), Dharmasvamin (4.697ab), Kalyanasvamin 
(6.697cd), and Mammasvamin (4.698-699). 

Kalhana reports only one Saiva foundation by a king of this dynasty, and this 
is a special case. For it was not the creation of a new Siva with the king's name, 
but merely the building by Lalitaditya of a new stone temple to house the ancient 
Siva Jyesthesvara at the site of Siva Bhutesvara (4.190) in the context of offer- 
ings to clear his debt to the latter incurred when he had appropriated the wealth 
of this temple to finance his military campaigns (4.189). Devotion to Visnu was 
also the preference of Avantivarman (r. 855/6-883), the first king of the next dy- 
nasty, and in keeping with his personal faith he installed an Avantisvamin before 
his consecration. But thereafter he showed himself a Saiva in unison with the 
faith of his powerful minister Sura, establishing a Siva Avantisvara and mak- 
ing donations to the Sivas of the national Siva-temples, confessing to Sura his 
long-hidden devotion to Visnu only at death's door (5.43, 123-125). 61 

Vaisnavism gained ground again only towards the end of our period, and in 
subsequent centuries. 62 Before that happened, while it remained in the shadow 
of Saivism, it gave rise to a new literature of scriptural texts known collectively 
as the Pancaratra, that was probably composed in and around Kashmir. A form 
of Vaisnavism bearing this name is already mentioned in the Mahabharata. &z 
It is very probable, therefore, that it was in existence well before the Saiva 
Mantramarga. However, there is no evidence that this early Pancaratra had 
a Tantric ritual system of the kind that characterizes the Samhitas of the sur- 
viving corpus of Pancaratrika scripture. It is highly probable in my view that 
those texts are rather the product of a thorough reformation in which Vaisnavas 
followed the example of the already flourishing Saiva Mantramarga in order to 
provide themselves with a substantially new ritual system that would enable 
them to compete more effectively with their rivals. 64 

61 For the remains of Avantivarman's Avantisvamin and Avantisvara temples, both 
built at Avantipura, see Krishna Deva in EITA vol. 2, pt. 1, pp. 368-373; plates 
734-738 and 740-757. 

62 Vaisnavas who left their mark in the domains of the major Sastras, belles-lettres, 
and literary theory are few during our centuries. The shift in the fortunes of 
Vaisnavism is marked by the emergence of such influential religious leaders as 
Ramanuja (d. 1137), Madhva (probably 1238-1317), Nimbarka (thirteenth cen- 
tury), Visnusvamin (thirteenth century?), Vallabha, and Caitanya (both late fif- 
teenth century). For an excellent survey of the history of these Vaisnava traditions 
see Colas 2003. 

63 Mahabharata 12.322.24; 12. 326.100; 12.360.76;12.337.1; 12.370.59, 63, and 67. 

64 It was this tradition that was subsequently adapted in South India as the basis of 
texts such as the Isvarasamhita, Padmasamhita, and Paramesvarasamhita, whose 
purpose, absent in the earlier Samhitas, was to provide scriptural authority for a 
Pancaratrika system of temple-worship. 


Genesis and Development of Tantrism 

I am led to this conclusion by the convergence of various considerations. 
Firstly, the ritual system prescribed in the Pancaratra scriptures is remarkably 
close to that of the Saiva Mantramarga in its repertoire, consisting principally 
of Mandala initiation (diksd), regular worship comprising Nyasa, Puja, Japa 
and Homa, the periodic ritual of pavitrdropanam, special rites of Mantra- 
propitiation (mantrdrddhanam), and image-installation (pratisthd); and this 
proximity extends into the minute details of the procedures of these rituals and 
even to the production of Vaisnava versions of such eminently Saiva rites as the 
vetdlasddhanam. 65 

Secondly, I see no evidence that any of the surviving Pancaratra texts goes 
back as far the Saiva texts that they so closely resemble. Seven can be shown 
to be relatively old because they have been cited by authors of the tenth cen- 
tury or have come down to us in early Nepalese palm-leaf manuscripts. These 
are the Svdyambhuvapancardtra, the Devdmrtrapancardtra, the Vdsudevakalpa 
of the Mahdlaksmlsamhitd, the Jayottara, the Jaydkhya, the Sdtvata, and the 
Pauskara. Now, of these, three, namely the Jayottara, the Jaydkhya, and the 
Sdtvata, are very unlikely to have been produced before the ninth century, that 
is to say, at a time when the Saiva Mantramarga had been flourishing under 
widespread royal patronage for at least two centuries and had been existence in 
some form by a time no later than the middle of the sixth and perhaps as early as 
the middle of the fifth. For all three focus on the worship of a form of Vasudeva, 
called Vaikuntha in the Jaydkhya and Jayottara and Saktyatman or Saktisa in 
the Sdtvatasamhitd, in which the principal anthropomorphic face is flanked by 
the faces of Narasimha and Varaha, with a fourth face, that of the sage Kapila, 
at the rear. 66 Surviving stone and bronze images of this deity are numerous, 
but they are three-faced, lacking the face of Kapila at the rear, until the ninth 
century 67 

Thirdly, these early Pancaratra texts show clear signs of having drawn on 
Saiva sources. This is particularly obvious in the Svdyambhuvapancardtra, to 
which we have access in a single, incomplete Nepalese palm-leaf manuscript 
bearing a date of transcription that falls in A.D. 1026. 68 The principal Mantra of 

65 A vetdlasddhanam is taught in Jayottara 8.23-26b. 

66 Jaydkhyasamhitd 6.73c-64 (JS) (=Jayottara 1.20 [J]): dhydyec caturbhujam 
*vipra (JS :devam J) sahkhacakragadddharam || caturvaktram sunayanam 
sukdntam padmapdninam | vaikuntham *narasimhdsyam (JS:ndrasimham ca 
J) vdrdham kapUdnanam; Sdtvatasamhitd 12.9, 14c-15: saktlso 'py atha 
samcintyah pundarlkanibheksanah | icchdrupadharas caiva saumyah praha- 
sitdnanah \\ . . . ndrasimhena vaktrena bhavabhitivighdtakrt \\pusndti sarvabhutdni 
vdrdhendmrtdtmand \ kurute pascimasthena kdpUenopasamhrtim. 

67 See Sanderson 2005b, pp. 283-284, drawing on Siudmak 1994. 

68 Svdyambhuvapancardtra, exposure llb3: samvat 147 dsddhasukla ekddasydm 


The Saiva Age 

this text, which may well be the oldest of the seven, is the well-known Vaisnava 
Dvadasaksara OM NAMO BHAGAVATE VASUDEVAYA NAMAH. But the principal 
among its ancillary Mantras are five that it calls the Brahmas. These are mani- 
festly adapted from the venerable Saiva Mantras of that name. 69 

sukradine +++ naksatre Hikhitam (corr. : liksatam Cod.) iti 'Copied on Friday, un- 
der the asterism +++, on the eleventh Tithi of the bright half of the month Asadha 
in the [expired] year 147'. That the unstated era of this date is the Newari, 
which began on 20 October, 879, is confirmed by palaeographical comparison with 
other Nepalese manuscripts of the early eleventh century. I am very grateful 
to Dr. Diwakar Acharya for providing me with a digital copy of this manuscript 
and his own transcription, and also for the information that a second manus- 
ript of this text photographed by the NGMPP (B 237/16) is merely a copy of the 
first. The title Svayambhuvapancardtra appears nowhere in the surviving folios 
but is reconstructed here from the analytic equivalent seen in the colophon of the 
eighth Adhyaya: iti pahcardtre *svdyambhuve (corr. : svayambhuve Cod.) astamo 
<'>dhydya<h>. The other surviving Adhyaya colophons refer to the work simply 
as pancaratram or pancaratram mahajnanam . The meaning is 'the Pancaratra 
of the Self-born', i.e. 'the Pancaratra taught to Brahma'. The text is indeed 
instruction given in response to questions posed by Brahma. The instructor is 
Siva/Isvara. Exposure 3al-2 (the beginning): OM NAMO BHAGAVATE VASUDEVAYA 
|| . . .pranipatya haram deva<m> . . . stutvd ndmasahasrena brahma vacanam 
abravlt; exposure 4a2-3: *brahmano vacanam (em.:brahmacanam Cod.) srutvd 
Tsvara<h> *pratyabhdsata (em. -.pratyubhdsyate Cod.) | srnu brahma<n> prayat- 
nena visno<h> sthdpanam uttamam | pancardtramahdjndnam sarvasdfstresu] cot- 

The Devdmrtapancardtra, which is closely related textually to the 
Svayambhuvapancardtra and is probably dependent on it, survives in a sin- 
gle, undated Nepalese manuscript, probably of the twelfth century. Here too I am 
indebted to Dr. Diwakar Acharya, who provided me with a transcript that he has 
69 The five Vaisnava Brahmas are as follows (Svayambhuvapancardtra, ex- 
posure 10al-2): om narenarenarannatha nara yasman narottama 



+ + + naJmas te rudrarudrebhyah trtlya brahma \ om tatsamyogaya 

Brahmas after the first are evidently modelled on the Saiva Brahmas in the 
order (1) Vamadeva (vamadevaya namo jyesthaya namo rudraya namah 


Aghora (aghorebhyo 'tha ghorebhyo ghoraghoratarebhyas ca sarvatah 


brahma sivo me 'stu sada sivah). The first Brahma has nothing in common 
with the remaining Saiva Brahma, that of Sadyojata. 


Genesis and Development of Tantrism 

The Saiva prototypes are already found in the Atimarga of the Pancarthika 
Pasupatas. Indeed they constitute the whole Mantra-system of that tradition. 
However, it is clear that the Svayambhuvapancaratra has drawn them from the 
later tradition of the Mantramarga, because it goes on to teach the imposition on 
to the worshipper's body of the thirty-eight parts of these Mantras (kalanyasah), 
a Mantramargic feature, and under names specific to one Mantramargic tradi- 
tion, that of the Svacchandatantra, the principal scripture of the Mantrapitha. 70 

The Svayambhuvapancaratra survives only in this Nepalese manuscript. 
One might object, therefore, that it may be no more than a local oddity unrep- 
resentative of the mainstream tradition. That it is not can be argued, of course, 
only through evidence that the text was more widely known in the form of ref- 
erences to it, citations from it, or accounts of its contents in other works. This 
is a difficult test to apply in the case of the early Pancaratrika literature, since 
in stark contrast to the case of the Saiva scriptures, Pancaratrika commentarial 
works in which we could seek such evidence are almost completely absent un- 
til a much later period among the Srivaisnavas of the South, when the range 
of relevant sources had changed greatly. The only exception is the Spandapra- 
dipika of the Kashmirian Bhagavatotpala, probably of the tenth century 71 But 
that, though it cites a number of early Pancaratrika scriptural sources, does not 
cite this. However, there is evidence in a Saiva source that this Pancaratrika 
text was known and followed outside Nepal. For I propose that it is identi- 
cal with the Svayambhupancaratra that Somasambhu cites as his authority in 
his account of the procedures for the installation of an image of Visnu in the 
KriyakandakramavalT, 72 the highly influential work on the Saiddhantika Saiva 

70 Ibid., exposure 10a3-5: kaldnydsam caturthan tu | srsti vrrddhi mati laksmT medhd 
kdnti svadhd sthitd | rajo raksd rati pdlyd kdmd trsnd mati jnayd | avidhi kdya 
tdta ca bhrdmam mohani tathd | + + + + + + + sthdh ksudhd mrtyu jvarabhayd 
| nirvitis ca pratisthd ca | sdnti vidyd tathaiva ca | tard sutdrd taranl tdrayanti 
svatdranT | astatrihsa*kalopeta (em. -.kaldpetah Cod.) deary ah *samuddhrtah 
(corr. : samuddhrtdh Cod.). Cf., to emend the names, Svacchandatantra 1.53-59b 
USvacchandalalitabhairava IFI T. 507, p. 6; NAK MS 1-224, f.3v4-4rl, the latter 
with different kaldh of Isana) and Netratantra 22.26-34. 

71 I am aware of no reference to the Spandapradipikd or its author in any dated work. 
It is not possible, therefore, to fix a date before which this work must have been 
written, at least not a date earlier than that of its manuscripts. However, the fact 
that it quotes extensively from the Sakta Saiva literature current in Kashmir up to 
and including the Isvarapratyabhijndkdrikd of Utpaladeva (fl. c. 925-975) but not 
from any of the works of Abhinavagupta (fl. c. 975-1025) makes it unlikely that its 
author wrote after the latter. 

72 Verse 4.12ab in Brunner's edition (Somasambhupaddhati, Pt. 4, p. 297) (B), 
= verse 1668cd in the KSTS edition (Karmakdndakramdvali) (K), and folio 
71v2-3 in the Cambridge MS (KriyakandakramavalT) (C): svayambhu*pahcardtre 
(NK: pdncardtre B) ca sarvam etad udiritam. 


The Saiva Age 

rituals 73 that he composed in the eleventh century, probably in 1073, 74 while he 
held the office of abbot in the kingdom of the Kalacuris of Tripuri at the illus- 
trious Saiddhantika monastery of Golagi (golaglmathah), in the Rewa District of 
Madhya Pradesh. 75 

My conclusion that Somasambhu was referring to our Svayambhuvapanca- 
rdtra does not rest solely on the synonymity of the titles, both meaning 'The 
Pancaratra taught to Brahma', but also on the fact that the brief but detailed 
account of the ritual that Somasambhu attributes to the Svayambhupancaratra 
corresponds in its particulars to the coverage of the same topic found in the 
seventh Adhyaya of the text in our manuscript. I cannot demonstrate this 
in full detail here. But it should suffice to point out that the system that 
Somasambhu attributes to his Svayambhupancaratra features an unusual 
arrangement of three circuits of Mantra-deities that agrees exactly with that 
of our Svayambhuvapancaratra manuscript: nine on a lotus with eight petals 
(one at the centre and one on each of the petals), twelve in a circle with that 
lotus at its centre, and eight forming a circuit enclosing the whole. The twelve 
are the Visnumurtis, embodying each of the twelve syllables of the root-Mantra 
(mulamantrah); the outer eight are the eight weapons (astrani) held by the 
presiding deity; and the nine of the innermost circuit (garbhavaranam) are a 
set of ancillary Mantras: the Hrdaya at the centre surrounded by the Siras (E), 
the Sikha (S), the Kavaca (W), the Astra (N), the Gayatri (SE), the Savitri (NE), 
the Netra (SW), and the Pirigalastra (NW). 76 Since this arrangement is highly 

73 Of the various Paddhatis on the Saiddhantika rituals that have come down to us 
Somasambhu's was probably the most influential. Its impact can be seen in the 
major later works of this type, such as the Kriydkramadyotikd of Aghorasiva, the 
Jndnaratndvali of Jnanasiva, and the Siddhdntasekhara of Visvanatha, and in the 
fact that manuscripts of the text have survived throughout the subcontinent, in 
Kashmir, Nepal, and the South. There is also the fact that it alone achieved the 
distinction of being stripped of its human authorship to be passed off as scripture. 
For it was incorporated almost in its entirety in the Agnipurdna (SANDERSON in 
Brunner 1998, p. lix, fn. 81); and much of it was taken over in the late south-Indian 
Saiddhantika scriptures Cintyavisvasdddkhya and Uttarakdmika (Brunner 1998, 
p. lviii-lix). 

74 For a discussion of the date of Somasambhu's Paddhati see Sanderson 2007a, 
pp. 420-421, footnote 640. 

75 For the name Golagi and the location of the monastery see here p. 264. 

76 The relevant passage in the Svayambhuvapancaratra (exposure 5b3-5a2) is as fol- 
lows (with some restorations and emendations following the readings of a closely re- 
lated passage in the eleventh Adhyaya of the Devdmrtapancardtra [D]): *yajanam 
(em. D and here, exposure 8a3 : ++ nam Cod.) sampravaksydmi *divyam (D :devam 
Cod.) ndrdyanasya *tu (D:tuh Cod.) | tribhir dvaranaih ' k kdryam (em. :kdya 
Cod. :kdrd D) durlabham *tu surdsuraih (D :sasurdsuram Cod.) | madhye cakram 
*pratisthdpyam (em. : pratisthdydm Cod. ipratisthdpya D) *dvddasdram (corr. [D: 
arai<r> dvddasabhir yutam]:dvddasdna Cod. ) susobhanam | tanmadhye ka- 


Genesis and Development of Tantrism 

unusual, especially in its set of nine ancillaries, it is extremely unlikely that 
Somasambhu's Svayambhupancaratra is not the Svayambhuvapancaratra of 
the Nepalese manuscript. Since Somasambhu was a major figure and writing 
far from Nepal for a pan-Indian audience there are no grounds for considering 
this tradition to be a Nepalese aberration. 

Furthermore, while the ritual systems taught in the scriptures of the 
Pancaratra are generally coherent, no less so than those of the Saivas, the texts 
retain elements that make sense in the Saiva world but not in the Vaisnava; 

malam proktam patrdstakasakarnikam | sarvdtmd *sakalo (em. :sakald Cod.) *devo 
(corr. : deva Cod.) *divyamdldsasamanvitah (conj. : divyamdldsandtanah Cod.) | 
sriyd madhye tu hrdayam humkdrena tu pujayet | sira<h> purvadale *dadydd 
daksine tu sikham (D:da + + + + + + + m Cod.) nyaset | pastime kavacam *dadydd 
(corr. : dadydv Cod.) astrah caivottarena tu | gdyatry dgneyadig*bhdge (corr. : bhdga 
Cod.) savitrlm Tsvare svayam | *netrah (corr. : netrdh Cod.) caiva tu *nairrtydm 
(corr. : nairitydm Cod.) pihgalastram tu *vdyave (corr. : vdyavet Cod.) | guhydd 
guhyataram guhyam garbhavaranam uttamam | *dvitlyam (corr. -.dvitiydm Cod.) 
*sampravaksyami (corr. : sampravaksyamih Cod.) visnu*murtlh (corr. :murtti Cod.) 
prapujayet | dvadasare tathd cakre nyase<d> dvadasa murtayah | *kesavam tu are 
purve omkarena (D:ke ++++++++++ rena Cod.) tu pujayet \ dvitiyan tu nakdrena 
*pujya (conj. :j 'hey 'dm Cod.) ndrdyanan Hatha (corr. : tathdh Cod.) | trtlyam 
mddhavam *pujya (em. : pujyam Cod.) mokdrena *mahdtmand (D : mahdtmanah 
Cod.) | bhakdrdksaradevena govindan tu *caturthakam (D : caturthakaih Cod.) | 
pancaman tu gakdrena visnu<m> caiva prapujayet \ vakdraksaradevena sasthe 
vai madhusudanam \ saptame vdmanah *caiva (corr. : caivah Cod.) tekdrena 
tu pujayeft] | *yajed vdkdrabljena (conj.:+j . dvdrabljena Cod.) astame tu 
Hrivikramam (corr. -.trivikramah Cod.) | srldharan navamah caiva sukdrena 
tu pujayet | dasame tu hrslkesam dekdrena tu pujayet | ekddase tu *vdkdre 
(conj. : vdkdra Cod.) padmandbham *prabhum (corr. prabhu Cod.) viduh 
dvddase <tu> bhakdrena ndmnd ddmodaram smrtam | *dvitiydvaranam khydtam 
(D : dvitlydvarana khydtam Cod.) Hrtlye 'strdni (D:trtiyena strdni Cod.) vinyaset 
| sahkha<m> caiva nyase<t> *purve (em. \purvvam Cod.) *dgneyydm tu gaddm 
nyaset (D:dgneyd +++++ Cod.) | *daksinena (corr. : + ksinena Cod.) bhave<c> 
cakram khadgam *nairrtyagocare (corr. : nairityagocaret Cod.) | padma<m> 
pascimato vidyd<d> vdyavydm tu hala<m> nyaset | musala<m> *cottarato (em. 
in spite of the metre :cottato Cod. D) dadydd Tsdnyd<m> *sdrhga (corr. : sdrdhga 
Cod.) vinyaset \ etad guhyataram *ydgam (corr. :ydgdm Cod.) durlabham para- 
mam padam. Somasambhu sets out the same material in his Paddhati in 4.27c-33 
of Brunner's edition, =vv. 1681c-1686 in the Kashmirian edition, and f. 72r2-7 
in the Cambridge manuscript (the last two sources offer no significant variants 
but only minor errors and corruptions that I have not recorded here): vinyasya 
cdditas cakram dvddasdram subhdsvaram || 28 tasya madhye punar deyam pad- 
mam astadalam tatah \ hrnmantram karnikdydm ca sirah purvadale tatah \\ 29 
sikham ca daksine patre pastime kavacam nyaset | astram uttarato nyasya gdyatrim 
agnipatrake || 30 savitrlm Tsapatre ca netram ca nairrte dale | tatas ca vdyupatre 
ca pihgalastram viniksipet || 31 garbhavaranam ity uktam adhundvarandntaram | 
dvadasare ca cakre 'smin kesavddydn yathdkramam \\ 32 pranavddyair yathdkdram 
uktapurvaih svandmabhih | prdgdditas ca vinyasya khadgam gaddm anantaram || 
33 cakram sahkham ca padmam ca halam ca musalam tatah | sdrhgam ca vinyased 
evam trtiydvaranam bhavet. 


The Saiva Age 

and in some cases we find a degree of awkwardness that is consistent only with 
a clumsy attempt to adapt Saiva materials to their new context. 

A striking example of this can be seen in the Jayakhya. When detailing the 
process of initation it describes the pdsasutram, the cord which is ritually trans- 
formed into a substitute of the subtle body of the candidate, containing all the 
reality-levels along its length, to be used in the process of rendering the past ac- 
tions that bind his soul incapable of giving rise to future consequences at any of 
these levels. In the course of this description we find some elements alien to the 
Vaisnava tradition that derive, with minimal distortion, from the Saiva doctrinal 
context. Thus it speaks of this cord as embodying kald, avidyd, and ragah, and, 
shortly afterwards, as coloured by ragah, illuminated by avidyd, circumscribed 
by kdlah, and rendered non-pervasive by niyatih. 11 Now the first three of these 
factors {ragah, avidyd, and kald) are the Saiva Mantramarga's three 'shrouds' 


The only edition of the Jayakhya (Ed.), that of KRISHNAMACHARYYA, was based 
on south-Indian manuscripts of relatively recent date. I re-edit the text of the 
passage to which I am referring, 16.128c-134 [numeration of Ed.], with the help 
of the testimony of a Nepalese paper manuscript of 1454/5 (N), ff. 35v7-36r4, 
and a lemma in a Nepalese palm-leaf manuscript of 1187/8 of the Jndnalaksmi 
of Sadhaka Candradatta, pupil of Ekayanacarya Narayanagarbha (C): susitam 
sutram adaya laksalaktakabhavitam || 129 sammukham cotthitam sisyam 
*samapadasirodharam (corr. [=C] : semapadasirodharam N : samapadasironnatam 
Ed.) | krtvahgusthadvayasyagrat samarabhya *dvijottama (Ed. : dvijottamah 
N) || 130 yavac chikhavasanam tu sutra*mdnam (Ed. : mana N) samaharet 
| kuryad *ekagunam (Ed. : vekagunam N) tad *vai (Ed. :ve N) dvigunam 
trigunam tu va || 131 *tris tris tad (conj. : tristrismad N -.tritristha Ed.) 
gunitam vatha *pancavimsatidhdthava (N -.pahcavimsati cathava Ed.) | avyak- 
talihgasutram tu Had ragavidyakalatmakam (em. : tadrdgrdvidydkaldtmakam 
N :pragavidyakalatmakam Ed.) || 132 *nityam jadam (Ed. :nityajade N) 
vyapakam ca tasmin visvam pratisthitam | Hatraivastam vrajed (corr. : tatrevastam 
vrajed N -.tatraptam ayate ~Ed.:tatrastam ayate conj. KRISHNAMACHARYYA) 
bhuyas tasmad eva pravartate \\ 133 tatrastham cintayet sarvdra abhinnam 
tattvapaddhatim | Hattvodbhavas (N :tatrodbhavas Ed.) tu ye vipra *pasa 
(em. :pasa Ed. : tesam Ed.) bandhatmaka drdhah || 134 ragena rahjitas *citra 
(Ed : cimta N) avidyasampradlpitah | vicchinnas caiva kalena *niyatyavyapakas 
(conj. : niyatavyapakas N Ed.) tatha 'O best of brahmins, after taking up a perfectly 
white cord soaked [red] with lac and making the candidate stand facing him with 
his feet together and his head upright, he should measure out [a length of] the cord 
from the tip of his two big toes to his hair-tuft. He may make [the cord of this 
length] single, double, triple, thrice triple, or twenty-fivefold. He should meditate 
upon the entire sequence of Tattvas as residing undivided therein. This thread, 
[which embodies] the subtle body [of the candidate], comprises Raga, Avidya, and 
Kala (ragavidyakalatmakam). It is eternal, unconscious, and pervasive. The whole 
universe is grounded in it. Into it it disappears again and from it alone it comes 
forth. These binding cords are the firm fetters [of the soul]. They arise, O brahmin, 
from the Tattvas. They are coloured because they have been dyed with [the red- 
ness of] Raga. They are illuminated by Avidya, circumscribed by Kala, and made 
non-pervasive by Niyati'. 


Genesis and Development of Tantrism 

{kancukani), except that there the second is generally termed vidyd rather than 
avidya; and the other two factors, kdlah and niyatih join these three to form the 
group of five reality-levels (tattvdni) ranked immediately below mdydtattvam, 
the upper limit and source of the 'impure cosmos' (asuddho 'dhvd), and immedi- 
ately above the individual soul (purusah), constituting the factors that enable 
the soul to undergo embodiment in that impure world. 78 Even the substitu- 
tion of avidya for the Saivas' vidyd does nothing to dilute the obviously Saiva 
character of the set, since vidyd in that context is indeed a form of nescience 
(avidya), being understood as the limited power of knowledge that characterizes 
bound souls, enabling them to cognize the objects presented by the faculties, as 
opposed to the pure, all-encompassing knowledge (suddhavidya) that operates 
above mdydtattvam; and this understanding is maintained in the passage in the 
Jaydkhya, because it speaks of the bonds as being 'illuminated' by avidya. Indeed 
the line in which the bonds are said to be 'coloured by ragah and illuminated by 
avidya' unmistakeably echoes loci classici on the functions of ragah and vidyd in 
the Mantramarga's scriptures. 79 

The Sdtvata and the Pauskara are probably the latest of these early texts. 
They are certainly the most polished and the most sophisticated in language. 
Unsurprisingly, these more mature products of the tradition contain no glar- 
ingly obvious examples that I can see of imperfectly assimilated Saiva material. 
Nonetheless, there are parallels in which the Saiva version seems more likely 
to have been the model of the Pancaratrika than vice versa. Thus the nine- 
teenth chapter of the Pauskara teaches as the text's major initiation Mandala 
(mahayagah) an arrangement of eight lotuses around a central ninth, calling it 
the navaplthamandalam, navabjamandalam, or navanabhamandalam, 80 and a 

78 For ragah, vidyd, and kala as the three 'shrouds' (kahcukatrayam) of the Saivas 
see, e.g., Matahgaparamesvara, Vidyapada 11.33: rdgavidydkaldkhyena kahcu- 
katritayena vai; and Rauravasutrasamgraha 1.3-4: rdgavidydkaldvyaktaguna- 
buddhisamudbhavam, where they are the three 'shrouds' (kancukani) of the 
bound soul. For the addition of kdlah and niyatih seen in the last verse 
of the Jaydkhya passage (16.134) see, e.g., Matahgaparamesvara, Vidyapada 
14.2: kahcukatritaydviddham kdlena kalitam sanaih \ niyatydlihgitam ydti 
pumbhdvendtmavartind; and Tantrdloka 9.204: mdyd kala rdgavidye kdlo niyatir 
eva ca \ kancukani sad uktdni. 

79 Cf. Svdyambhuvasutrasamgraha 32.10-11: kalodbalitacaitanyo vidyadarsita- 
gocarah | ragena ranjitas cdpi buddhyddikaranais tatah \\ mdyddyavani- 
paryantatattvabhutdtmavartmani | bhuhkte tatra sthito bhogdn bhogaikarasikah 
pumdn; Kirana 1.16c-17a: tayodbalitacaitanyo vidyakhyapitagocarah ragena 
ranjitas cdpi; and Kubjikdmata 13.3: ragena ranjitatma vai niyatyd yo 
niydmitah avidydprerito gacchet svargam vd svabhram eva vd. 

80 Pauskarasamhitd 1.24ab: yady ekam tu mahdydgam navandbham samudyajet; 
10.34cd: navapithe mahdydge tarn ca krtsnam vaddmi te; 19.26: yair uddistam 
mahdydge navdbje. 


The Saiva Age 

long invocatory Mantra consisting of eighty-one units distributed one by one on 
the centre (nabhih) and eight petals of each of the nine lotuses. This arrange- 
ment and correlation, which, to my knowledge, is found in the Pancaratrika liter- 
ature only in the Pauskara, is central to the Saiva tradition of the Mantramarga, 
being the hallmark of a number of its earlier scriptures, where the Mandala is 
taught under the same names, 81 and the Mantra with which it is correlated is the 
well-known Saiva Vyomavyapimantra of eighty-one units. In the Saiva case the 
nine lotus-thrones (pithah) of the Mandala are equated with nine Tattvas: Siva, 
Sadasiva, Isvara, Vidya, Maya, Kala, Niyati, Purusa, and Avyakta (Prakrti). In 
the Pauskara that element has been dropped, no Vaisnava set of nine Tattvas 
being available for this purpose and the Saiva set being unassimilable because it 
includes unmistakeably Saiva elements such as Sadasiva and Isvara. Nonethe- 
less the text contains a sign that the redactor was after all working with a Saiva 
exemplar. For he calls his fourth 'the lotus of Maya'. 82 Maya is a Saiva not a 
Pancaratrika Tattva. 

Furthermore, in the Pauskara, the Satvata, and the Vasudevakalpa of the 
Mahalaksmisamhita we find the term spandah 'vibrancy' in the sense it has 
in the Sakta Saiva Jayadrathayamala and the Spandakarika of Kallata in the 
second half of the ninth century. However, I do not exclude the possibility that in 
this case it may be the Saiva sources that are indebted to the Vaisnava. 83 

81 Matahgapdramesvara , Kriydpdda 1.51c: mandalam navapithdkhyam; Ksemaraja, 
Svacchandoddyota vol. 2 (Patala 5), p. 22: navandbham navandbhisthdnastha- 
padmam etat puramandalam. Cf. Nisvdsaguhya 11.14ab (Nisvdsatattvasamhitd 
f. 83vl): ekdsitipado ydgo navavyuheti *samjnitah (conj. \samsthitah Cod.). 

82 Pauskarasamhitd 19.24c— 26b, 27ab, 31ab, 37c-38b: jhatum icchdmi 
*vidydkhyamantrdndm (vidydkhya em. widydkhyam Ed.) laksanam vibho \\ 
25 yaih padmakalpand kdryd *padair (conj. :padmair Ed.) nirvartitaih prabho \ 
brahmaprakdsakdndm tu mantrdndm atha laksanam \\ 26 yair uddistam mahdydge 
navdbje pujanam tathd \ ... 27 madhyapadme paddndm ca navakam parikirtitam 
| ... 31 mdydmaye 'tha (conj. 'nte Cod.) kamale caturthe tu padam smrtam | . . . iti 
vidydpaddndm ca svarupena prakdsitam || 38 atha brahmapaddndm ca laksanam 

83 See Pauskarasamhitd 27.274—276: sdntasamvitsvarupasya spandananda- 
maydtmanah | tavdcyutam hi citspandam svayam parinatam 
smaret || 275 sahasrasasisurydgniprabhayd projjvalam sthiram | 
marlcicakrasampurnacidgarbham sarvatomukham || 276 cidambardntardvastham 
susdntam bhagavatpadam; Sdtvatasamhitd 3.15cd: evam jndtvd sthitim brdhmim 
svanandaspandalaksandm (conj. \svdnandam spandalaksandm Ed.); also 
5.99-101b: lolibhutam abhedena smaret turydtmand purd | nityoditam ca supade 
sthitam aspandalaksanam || 100 athdrcitum yam icchet tu visesavyaktilaksanam 
| samkalpya tu svabuddhyd tu tatkdlasamanantaram \\ 101 dhruvd sdmarthyasaktir 
vai spandatam eti ca svayam; Vasudevakalpa at 165ab: cicchaktau tu layam krtvd 
svdnandaspandagocare; 238-241b: mdnasena tu *ydgena (conj. :yogena draft 
Ed.) dravyaih samkalpajaih subhaih | hrdambujapare turye *cidbhdsdrupam 
(corr. : cidbhdsd rupam draft Ed.) uttamam || 239 kadambagolakdkdram 


Genesis and Development of Tantrism 

Nor was the influence of the Saivism of the Mantramarga confined to the 
formative period of the Tantric Pancaratra. For, as I have shown elsewhere, the 
Laksmltantra and Ahirbudhnyasamhita, works composed in the South, derive 
their distinctive doctrinal character from the assimilation of the dynamic non- 
dualism of the works of the Kashmirian Sakta Saivas from Utpaladeva (fl. c. A.D. 
925-975) to Ksemaraja (fl.c. 1000-1050). 84 

Royal Patronage of Buddhism 

Buddhism enjoyed widespread royal support during this period, notably 
from the Visnukundis of Andhra in the fifth and sixth centuries, from the 
Maitrakas of Valabhi in Saurastra in the sixth and seventh, from the Karkotas 
of Kashmir in the eighth, and throughout our period from the Licchavi and 
Thakurf kings of Nepal and various dynasties of eastern India, most notably 
the Palas (r. c. 750-1200). 

The Visnukundis of Andhra 

Among the eight successive Visnukundis (r. c. 375-612) known to us from in- 
scriptions three of the last six are known to have been patrons of Buddhism: the 
third, Govindavarman I (r. c. 422-462), the fifth, Vikramendravarman I (r. c. 502- 
527), and the seventh, Vikramendravarman II (r. c. 555-572). In the Tummala- 
gudem plates (Set I) issued by Maharaja Govindavarman I he is described as 
having beautified his kingdom with many temples and Buddhist monasteries, as 
having given generously to brahmins and Buddhist monks, as having resolved to 
attain the Great Awakening for the salvation of all living beings, and as having 
donated two villages — the charter's object is to record this grant — to fund the 

suryayutasamaprabham | svanandaspandarupam ca samcintyatmanam atmana 
|| 240 paranandasvabhavastho vetti yah pujanam vibhoh | tenarcitenarcitam vai 
dvisaptabhuvanatmakam || 241 visvam dyavaprthivl ca sadevasuramanusam; 
and 274c— 275: tanmadhye vistarastham ca laksmim sampujya purvatah 
|| vinyaset svasarlrac ca gurur vai pranayogatah | anandaspanda*rupam 
(corr. :rupam draft Ed.) capy amrtamrtarupinim. On spandah in Satkas 2-4 of the 
Jayadrathayamala see SANDERSON 2007a, pp. 365-366, 406, fn. 579. The term also 
occurs in the earlier first Satka, f. 190v4-5 (45.121-123b): nistarangarnavakarah 
paritrpta<h> paraparah \ susantamurtih sarvatma nirvaneso 'tinirmalah || 122 
tasya saktih svakam vlryam ciddhamanandagocaram | vyaktam vyaktivibhedena 
spandananandasundaram || 123 taddharmadharminl jneya saktir adya sivasya 
sd. For evidence that the first Satka of the Jayadrathayamala once formed an 
independent whole to which Satkas 2-4 were added in Kashmir at a later date see 
Sanderson 2002, pp. 2 and 22, n. 13, and 2005b, pp. 278-283. 
84 For the evidence see Sanderson 2001, pp. 35-38. For some other Saiva features in 
Paiicaratrika texts see Rastelli 2007, pp. 209, 214, and 224-225. 


The Saiva Age 

expenses of a Buddhist monastery founded by his chief queen Paramadevi. 85 A 
second set of plates discovered at Tummalagudem contains a charter issued by 
Vikramendravarman II which records his granting a village for the support of the 
Buddhist community at this monastery The founder's husband Govindavarman 
I is described as having beautified the whole of the Deccan with splendid Stupas 
and monasteries, and Vikramendravarman I, his grandson and the grandfather 
of Vikramendravarman II, is identified as paramasaugatah 'entirely devoted to 
the Buddha'. 86 However, in a charter issued by Vikramendravarman II in the 
previous year, recording a grant of a village to a Saiva temple, he is referred to 

85 SANKARANARAYANAN 1977:1, 11.8-24: anekadevdyatanavihdrasabhdprapdtaddko- 
dapdndrdmapratisamskdrdpurvakaranendlamkrtasakaladigantarena bhiksu- 
dhanasamudayendsakrd asakrt svasarvasvatydgind . . . sakalasattvadhdtutrdnd- 
yotpdditamahdbodhicittena mahdrdjasrlgovindav armaria . . . svasyd agramahisydh 
paramadevyd vihdrasya dipadhupagandhapuspadhvajapdnabhojanasayandsana- 
gldnabhaisajyakhandasphutitasirnasamskdrddikusalamuldnuccheddrtham dvdv 
ermad[d]laprenkaparundmadheyau grdmau udakaddnapurvakam atisrstau 'In 
order that his roots of merit should not be cut off, through [the provision of funds 
for] such [expenses] as lamps, incense, scents, flowers, banners, drinking water, 
food, beds, seats, medicines for sick [monks], and repairs to whatever is broken, 
cracked, and delapidated, the two villages named Ermadala and Preftkaparu 
have been donated to the monastery of his chief queen Paramadevi with the [due] 
pouring of water [into the hand of the recipient] by Maharaja Govindavarman, 
who has adorned all parts [of his kingdom] through his unprecedented provision 
of numerous temples, Buddhist monasteries, meeting halls, fountains, reservoirs, 
wells, and gardens, all of whose great wealth, lawfully acquired, is being enjoyed by 
Buddhist monks, brahmins, the unprotected, supplicants, the sick, the wretched, 
and the poor, who has [in this way] repeatedly given away all his property, and who 
has generated the intention to attain the Great Awakening for the salvation of all 
living beings'. 

86 SANKARANARAYANAN 1977:8, 11. 10—18: paramasaugatasya mahdrdjasrivikrame- 
ndrasya sunor . . . sri-indrabhattdrakavarmanah priyasunus . . . sri[md]n vikrame- 
ndrabhattdrakavarmd . . . ittham avabodhayati 'Vikramendrabhattarakavarman, 
beloved son of Indrabhattarakavarman, the son of paramasaugatah Maharaja 
Vikramendra informs you as follows'; 11.24—33: atibahuprakdramanoramo- 
ddrakarmddbhutastupavihdraculdmanibhir alamkrtasakaladaksindpathasya 
. . . srlgo[vi]ndardjasya murtimatlm sriyam praty avisayikrtamanorathayd para- 
ma[bha]ttdrikdmahddevyd srlmadindrapuram uccair alamkartukdmayeva prati- 
sthdpite srimati paramabhattdrikdmahdvihdre 'smdbhifh] . . . cdturdasdryavara- 
bhiksusamghaparibhogdya . . . irundoro ndma grdmo dattah 'I have donated the 
village called Irundora for the use of the community of excellent monks of the 
four directions in the venerable Paramabhattarikamahavihara that was founded 
by Paramabhattarikamahadevi as though desiring to bestow great beauty on 
Indrapura, fulfilling [thereby] the desire for embodied [royal] splendour of [her 
husband] King Govinda, who adorned the whole of the Deccan with splendid Stupas 
and monasteries that were marvelous in their most various, charming, and noble 


Genesis and Development of Tantrism 

as paramamahesvarah, as is his father Indrabhattarakavarman, 87 drawing to 
our attention that if a king supported Buddhism he did not necessarily cease to 
support other faiths or abandon his own. 

The Maitrakas ofValabhl 

Of the land-grant documents of the Maitrakas ofValabhl three quarters are 
records of grants to brahmins, but the remaining quarter report grants made by 
these kings to Buddhist institutions. 88 Guhasena (r. c. 553-569) has the epithet 
paramopasakah 'devout lay Buddhist'; 89 Slladitya I Dharmaditya (r. c. 595-612) 
is praised for his support of Buddhism in the east-Indian Rajavyakarana of the 
Buddhist Tantric Mafijusriyamulakalpa 90 and by the Chinese Huili in his ac- 
count of the Indian travels of Xuanzang; 91 and the latter, who visited the king- 
dom ofValabhl in the 630s, when the Maitraka Dhruvasena II was on the throne, 
reports that the king had recently developed a sincere faith in Buddhism and be- 
come a generous donor to the monastic community 92 Moreover, Valabhi became 
a major centre of Mahayana Buddhist scholarship during this period, producing 
such eminent figures as Sthiramati (fl. c. 510-570), for whom a monastery was es- 
tablished in Valabhi during the reign of Guhasena. 93 In their inscriptions, how- 

87 The Chikkula plates of Vikramendravarman (Sankaranarayanan 1977:7), 11. 15- 
19: parama[md]hesvarasya mahdrdjasya sri-indrabhattdrakavarmana[h] priyajye- 
sthaputro . . . paramamdhesvaro mahdrdjafh] srimdn vikramendravarmd evam d- 

88 SCHMIEDCHEN 2007, p. 360. 

89 SCHMIEDCHEN 1993, p. 84. 

90 Mafijusriyamulakalpa 53.537d-540: samudratiraparyantam laddndm jana- 
pade tathd || 38 sildhvo ndma nrpatih buddhdndm sdsane ratah | purim 
valabhya samprdpto dharmardjd bhavisyati || 39 vihdrdn dhdtuvardn citrdn 
*sreyase (em. : sreyasdm Ed.) prdnindms tathd | kdrayisyati yuktdtmd bhupatir 
dharmavatsalah || 40 pujdm ca vividhdkdrdm jinabimbdm manoramdm | pujayed 
dhdtuvardn agrydn lokandthebhyo yasasvisu \ ndsau mantrasiddhas tu kevalam 
karmajottamah 'In the land of the Latas up to the shore of the [western] ocean a 
king called Sila, devoted to the teaching of the Buddhas, will become a Dharma- 
raja in the city of Valabhi. That royal friend of Buddhism, of well-disciplined 
mind, will build monasteries and beautiful relic Stupas for the welfare of living 
beings. [He will establish] the manifold worship of beautiful images of the Bud- 
dha; and he will venerate the most excellent of the relics of the renowned Buddhas. 
He will not achieve success through [the Buddhist Way of] Mantras, but will ex- 
cell simply through acts of [lay] piety'. For the east-Indian origin of the text see 
Mafijusriyamulakalpa 53.627a: gaudadese 'smin; and 53.810a: prdcyadese 'smin. 

91 Beal 1914, p. 148. 

92 Xiyu ji, vol. 2, pp. 267-268. For a detailed account and analysis of religious patron- 
age under the Maitrakas during the sixth and seventh centuries see NJAMMASCH 
2001, pp. 199-278. 

93 On the dates of Sthiramati and the evidence that a monastery was established for 
him see Frauwallner 1961, pp. 136 ff. 


The Saiva Age 

ever, Siladitya I Dharmaditya, Dhruvasena II, and generally Guhasena too, ap- 
pear like almost all the other Maitrakas with the epithet paramamahesvarah. 94 

The Karkotas of Kashmir 

No inscriptions have survived from the reigns of the kings of the Karkota 
dynasty of Kashmir. But from the account of this dynasty given by the Kashmi- 
ran historian Kalhana we learn that although, as we have seen, the temples they 
established with their names were Visnus, 95 they or those closely associated with 
them also established several Buddhist foundations: the Anantabhavanavihara 
founded by the queen of Durlabhavardhana (r. c. 626-662); the Prakasikavihara 
founded by Prakasadevi, queen of Candrapida (r. c. 712-720/1); the Rajavihara 
'The King's Monastery' founded and richly endowed by Lalitaditya (r. c. 725- 
761/2) with a large Caitya and a huge Buddha image at his new capital 
Parihasapura; the Kayyavihara, founded during the rule of the same by Kayya, 
a king of Lata; a Vihara, a Stupa, and golden Buddha images established at 
Parihasapura by Lalitaditya's Central Asian chief minister Cankuna; a Vihara 
and a Caitya established by the same in the capital; and a large monastery 
and three Buddha images established by Jayapida (r. c. 773/4-804/5) in his new 
capital Jayapura. 96 

94 See, e.g., the Alma copper-plate inscription of Siladitya VII of a.d. 766/7, CII.39. 
There all the kings listed are said to be paramamahesvarah: the general Bhatarka, 
the founder of the dynasty, followed, after an unspecified number of generations, by 
Guhasena, Dharasena (II), Siladitya (I), Kharagraha (I), Dharasena (III), Dhru- 
vasena (II), Dharasena (IV), Dhruvasena (III), Kharagraha (II), Siladitya (II), 
Siladitya (III), Siladitya (IV), Siladitya (V), and Siladitya (VI). In the Maliya 
copper-plate inscription of Dharasena II, A.D. 571/2, we are given the names of 
the Maitrakas who ruled between the founder Bhatarka and Dharasena II. They 
are Dharasena I, Dronasimha, Dhruvasena I, and Dharapatta. Of these the first 
two have the epithet paramamahesvarah; Dhruvasena is here a Vaisnava (param- 
abhagavatah) rather than a Buddhist (paramopasakah); and Dharapatta is a devo- 
tee of the Sun-God (paramadityabhaktah . It seems that in the later years of the 
Maitraka dynasty, when Saivism had become firmly established as the religion 
of this dynasty, there was a desire to forget those early rulers, Dhruvasena and 
Dharapatta, whose religious preference had deviated. This practice of beginning 
the account of lineage with Bhatarka and then jumping to Guhasena and his suc- 
cessors, so that all the kings have the epithet paramamahesvarah, is already seen 
in the Dana plates of Dhruvasena II issued in 634/5 (EI 42:15). 

95 See here, p. 60. 

96 Rajatarahgini 4.3 (Anangabhavana); 4.79 (Prakasikavihara); 4.200-205 
(Rajavihara etc.); 4. 210 (Kayyavihara); 4.211 and 215 (the foundations of 
Cankuna); and 4.507 (the foundations of Jayapida). For the vestiges of Lalitaditya's 
Rajavihara, his Caitya, and Cankuna's Stupa at Parihasapura (Paraspor) see 
Krishna Deva in EITA vol. 2, pt. 1, pp. 366-367; plates 722-727. Cankuna is 
evidently a rendering of the Chinese military title jiangjun 'General' rather than a 


Genesis and Development of Tantrism 

The Licchavis of Nepal 

In the Kathmandu valley the inscriptions of kings throughout our period 
show their devotion to Siva. But here too, where Buddhism and Saivism 
co-existed among the Newars down to the present, there is ample evidence 
of royal support for the former. The Licchavi Vrsadeva is described in an 
inscription of his eighth-century descendant Jayadeva as having inclined to- 
wards Buddhism; 97 a view confirmed by a local chronicle, which attributes to 
him the establishing of Buddhist images; 98 and in the first half of the seventh 
century Xuanzang claims that the king of Nepal was a sincere believer. 99 The 
Gopalarajavamsavali, the earliest of the local chronicles, compiled during the 
reign of Jayasthitimalla (1382-1395), 10 ° claims that the Caitya at Gumvihara 
and a monastery, the Manavihara, were established by Manadeva, the Caitya 
of the Sinagu-vihara (the Svayambhunath Caitya) by Vrsadeva, 101 the Dhar- 
madevacaitya (the Cabahil Caitya) by Dharmadeva, a monastery and the 
Khasaucaitya (the Bodhnath Caitya) 102 by Sivadeva, the Phutovihara and a 
Caitya by Campadeva, the Rajavihara by Amsuvarman, the Devalavihara by 
Devaladeva, and a monastery at Nandisala by Sivadeva. To Narendradeva and 
his Buddhist preceptor Bandhudatta it attributes the instituting of the annual 
chariot festival (yatra) of the popular Newar Buddhist deity Bugmalokesvara 

97 LKA 148, 1.9: sugatasasanapaksapati . 

98 Levi 1990, vol. 2, p. 98. 

99 Xiyuji, vo\.2, p. 81. 

ioo rp^g Qopalardjavamsdvali, preserved in a single, palm-leaf manuscript that has 
lost the first sixteen of its folios, consists of three originally separate parts. The 
first (ff. 17r-30v) covers the period down to 1386. Its coverage of the period before 
the reign of Anantamalla (1274-1307) (ff. 17r-26r) consists of little more than a 
list of kings, the lengths of their reigns, in some cases a record of their religious 
foundations and a few contemporary events such as plagues and famines and rituals 
undertaken to avert them. From f. 26v to f. 29r it is a little more forthcoming. The 
last event it records is dated in 1379. Up to this point the text is in a low register 
of Sanskrit. The remainder of the first part, f. 29v-30v, is written in Old Newari in 
a more annalistic style and extends the account down to 1386. The second text (ff. 
30v-36r), in Old Newari mixed with Sanskrit, covers the years 1056/7 to 1275/6. It 
consists for the most part of chronological genealogy, giving dates of birth, length of 
reign, and age at death. The third (ff. 36v-63v + another f. 50), in Old Newari, is 
an annalistic chronicle whose main concern is to record religious foundations, with 
entries extending from 1258/9 to 1388/9. See Petech 1984, p. 6. 

101 The manuscript gives the name Visvadeva here, but as the editors propose, this 
is surely an error for Vrsadeva (f. 20r2-3): raja srivisvadeva varsa 100 tena krta 
sinaguvihara caityabhattarike pratisthita sampurna krtam. The identification of 
this with the famous Svayambhunath Caitya is evident from the name Sinagu, 
which corresponds to Syangu, its modern Newari name. 

102 This identification follows from the fact that the Bodhnath Stupa is known as Khasa 
Caitya in Newari. On these early Nepalese Caityas — this term rather than Stupa 
is the normal uage in Nepal — see Gutschow 1997, pp. 85-99. 


The Saiva Age 

(Bugadyah/Karunamaya-Matsyendranath). 103 Unsurprisingly, the Amaravati- 
mahavihara (Buga Bahah) at Bungamati, the home of Bugmalokesvara, claims 
to have been founded by him. 104 

Manadeva's dated inscriptions range in date from 459 to 505/6, 105 and 
we know from his Carigunarayana inscription that Vrsadeva was his great- 
grandfather and Dharmadeva his father. 106 The claim that he founded a 
monastery with his own name, the Manavihara, is confirmed by its mention 
in an undated inscription assigned to his reign. 107 The epigraphical dates of 
Sivadeva range from 590/1 to 604/5. 108 There is another Licchavi with the same 
name, with inscriptions ranging from 694 to 705, 109 but it is unlikely that it 
is the second that is intended, since grants of villages to the Sivadevavihara 
have been mentioned in two inscriptions dated in 679, during the reign of his 
predecessor. 110 The inscriptions of Amsuvarman range from 593 to 615; 111 and 

103 Gopdlardjavamsdvali f. 20v5: Caitya at Gumvihara; f. 21rl: Manavihara; f. 20v2- 
3: Caitya at Svayambhu; f. 21r3: Dharmadevacaitya; f. 21vl: Khasaucaitya; 
f. 21v2: Phutovihara and Caitya; f. 22vl: Amsuvarman's Rajavihara; f. 22v3: 
Devalavihara; f. 22v5: Sivadeva's monastery; and ff. 22v5-23rl (the festival 
of Bugadyah): sri narendradeva varsa 35 tasya dcdryabamdhudattadvayena 
srTbugmalokesvarabhatdrakasyaja.tra.krtd bhavati 'Narendradeva: [reigned for] 35 
years. Jointly with his Acarya Bandhudatta he established the festival of Lord Bu- 
gmalokesvara'. On the festival of Bugadyah, also known (in Nepali) as Rato ('Red') 
Matsyendranath, which is still a major event in the Kathmandu valley, see Locke 
1980, pp. 244-280. 

104 See the tabulated list of the eighteen principal monasteries of Patan and their 
founders in Locke 1980, pp. 32-33. He includes the Buga Bahah at its end, noting 
that it stands apart, not being counted among the principal monasteries of either 
Patan or Kathmandu. 

105 In the Licchavi inscriptions of LKA the earliest date is 464/5 (no. 2) and the latest 
505/6 (no. 19). An earlier inscription, dated in Vaisakha 381 (=a.d. 459), which 
came to light during renovation work at the Pasupati temple, has been published 
(Dhakal 1990). The earliest Licchavi dates are in the Saka era, which was used 
until the time of Amsuvarman, the last recorded Saka date being 526 (a.d. 604/5) 
in LKA 69 and 70. Thereafter the inscriptions are dated in a new era, often called 
Amsuvarman's, which commenced in A.D. 576, and continued in use until the intro- 
duction of a new era in Kartika 879, which has remained in use down to modern 

106 LKA 2, side 1, 1. 8-side 2, 1. 3: rdjdbhud vrsadevah . . . yasydbhut tanayah . . . raja 
sahkaradeva ity anupafmo] . . . devT rdjyavati tu tasya nrpater bhdryd . . .yasydm 
jdta . . . srlmdnadevo nrpah. 

107 LKA 18, 1. 18: ksetram cdksayam dattam [sri]mdnavihdre. 

108 LKA 54 and 70.' 

109 LKA 138 and 143. 

110 LKA 133, 11.4-11 and 134, 11.4-12: ayam grdmo . . . srisivadevavihdfre] catur- 
disdryabhiksusahghdydsmdbhir atisrstah 'I have given this village to the congre- 
gation of noble monks of the four directions at the Sivadevavihara'. 

111 LKA 59 and 85. 


Genesis and Development of Tantrism 

the Rajavihara arributed to him by the chronicle is mentioned in one of these, 
dated in 608. It also mentions the Manavihara and the Gumvihara, showing 
the accuracy of the report of the chronicle that these three monasteries are 
ancient Licchavi foundations. Moreover, it does so in a context that enables us 
to gauge their relative importance. For it fixes cash allowances from the court 
(rajakulam) to a large number of religious foundations and these are ranked into 
two groups. The upper comprises the temple of Bhagavat Pasupati, the national 
Siva, to whom all Nepalese kings from the time of Amsuvarman onwards have 
declared their allegiance, 112 Dolasikharasvamin (Carigunarayana), the principal 
Visnu of Nepal, then these three Buddhist monasteries, and two others not 
mentioned by the chronicle, the Kharjurikavihara and the Madhyamavihara. All 
of these are to receive the same allowance; and this is twice that to be received 
by the institutions listed in the lower group. That comprises "the ordinary 
monasteries" and the temples of various other deities, most of whom are Sivas, 
including Manesvara, evidently the temple of a Liriga installed by Manadeva 
with his name. 113 Narendra, whom the chronicle reports to have instituted the 
annual chariot festival of Bugmalokesvara, has dated inscriptions from 643 to 
679. 114 The last two, issued in 679 and mentioned above for their reference 
to the Sivadevavihara, record the granting of villages to that monastery; and 
the Chinese envoy Wang Xuan-ce reported that when he had an audience with 

112 See Sanderson 2005a, p. 417, fn. 254. 

113 LKA 77, 11. 6-15: bhagavatah pasupateh pu 6 pa 2 dolasikharasvaminah pu 6 pa 
2 +++ gumviharasya 6 pa 2 srimanaviharasya pu 6 pa 2 srirajaviharasya 6 pa 2 
kharjurikaviharasya 6 pa 2 ma[dhya]maviharasya 6 pa 2 samanyaviharanam pu 3 
pa 1 ramesvarasya pu 3 pa 1 hamsagrhesvarasya pu 3 pa 1 manesvarasya pu 3 pa 
1 sambapurasya pu 3 pa 1 vagmatiparadevasya pu 3 pa 1 dharamanesvarasya pu 3 
pa 1 parvatesvarasya pu 3 pa 1 narasimhadevasya pu 3 pa 1 kailasesvarasya pu 3 
pa 1 bhumbhukkikajalasayanasya pu 3 pa 1 tadanyadevakulanam pu 2 pa 2 . . . 'six 
Pufranas] and 2 Pafnas] each for Bhagavat Pasupati, Dolasikharasvamin (=Cangu- 
narayana), the Gumvihara, the Manavihara, the Rajavihara, the Kharjurikavihara, 
and the Madhyamavihara; 3 Pufranas] and 1 Pa[n]a each for the ordinary Viharas, 
and [the temples of Siva] Ramesvara, the Lord of the Hamsagrha (=Visnu Lokapala- 
svamin), [Siva] Manesvara, Samba[siva], Vagmatiparadeva [Siva], [Siva] Dhara- 
manesvara, [Siva] Parvatesvara, Narasimhadeva, [Siva] Kailasesvara, and the 
[Visnu] Jalasayana of Bhumbhukkika (=the Visnu of Budhamlkanth); 2 Pu[ranas] 
and 2 Pafnas] for the temples other than these . . . '. The Kharjurikavihara calls to 
mind the Stupa which the Buddha predicts in the Mulasarvastivadavinaya will 
be built by the Kusana emperor Kaniska at Kharjurika four hundred years af- 
ter his Parinirvana (Gilgit Manuscripts, vol. 3, pt. 1, pp. 1, 1.20-2, 1.5: bhagavdn 
kharjurikam anupraptah | . . . esa caturvarsasataparinirvrtasya mama vajrapane 
kanisko nama raja bhavisyati | so 'smin pradese stiipam pratisthapayati | tasya 
kaniskastupa iti samjna bhavisyati. 

114 LKA 123-134. 


The Saiva Age 

Narendradeva in 643 the king's belt was adorned with a Buddha. 115 But here 
too we see that the support of Buddhism in Nepal as elsewhere was not a sign 
that a king had changed his religious allegiance in any radical sense. For in both 
of those inscriptions Narendradeva has the epithet paramamahesvarah. 116 

The Thakuri Kings of Nepal 

Between the Licchavis, who last appear in the epigraphical record in 737, 
and the Malla kings, who ruled from 1200-1768, lies the relatively obscure 
period of the so-called Thakuri kings. These too, though predominantly Saiva, 
supported Buddhist institutions. Only one, Simhadeva (r. 1110-1126), has been 
declared paramasaugatah; 111 but several of the monasteries of the Kathmandu 
valley are attributed to kings of this period in inscriptions, palm-leaf deeds, 
manuscript colophons, or their own tradition: the Padmacakramahavihara to 
Gunakamadeva I, 118 the Jyotirmahavihara (Jyo Bahah) and Dattamahavihara 

115 The report of this encounter has been incorporated in chapter 221 of the Jiu Tang- 
shu (Old History of the Tang Dynasty), covering the years 618-906 and compiled in 
940-945. In a translation of this passage published by Sylvain Levi (1894, p. 67) 
we read "Leur roi Na-ling ti-po (Narendra Deva) . . . a . . . des breloques a sa ceinture, 
ornees d'un Fou-tou (Buddha?)". In a footnote he explains the question mark, say- 
ing that the use of fou-tou for 'Buddha' in the seventh century is problematic. But 
when he re-published his translation (1905a, vol. 1, p. 164) he removed the question 

116 LKA 133, 11. 1-3: bhagavatpasupatibhattdrakapdddnugrhlto bappapdddnudhyd- 
to licchavikulaketuh paramamdhesvaraparamabhattdrakamahdrdjddhirdjasrlna- 
rendradevah kusali gullahgahgrdmanivdsinah pradhdnapurahsardn sarvakutu- 
mbinah samdjndpayati 'Favoured by the venerable lord Pasupati, devoted to his 
venerable father, the banner of the Licchavi dynasty, entirely devoted to Siva, the 
supreme Lord, the paramount king Narendradeva greets the elders and all the 
other householders who live in Gullarigan village and commands them [as follows]'. 
The same formula is seen in 134, 11. 1-4. Only the name of the village differs. 

The historicity of Campadeva and Devaladeva, the remaining two kings men- 
tioned by the Gopdlardjavamsdvali as the founders of monasteries, is doubtful. 
They appear nowhere in the corpus of known Licchavi inscriptions, and in the 
local chronicles only in the Gopdlardjavamsdvali, which places the first between 
Sivadeva and Narendradeva and the second before Dhruvavarman — another name 
found only in this source — and Bhimarjunadeva. 

117 Colophon of ASB MS 9973 (Shastri 1917, pp. 4-5): paramasaugatasrimatsimha- 
devasya vijayardjye. 

118 Petech (1984, p. 40) quotes the following colophon of an Astasdhasrikd 
Prajndpdramitd MS (NAK 3-359) that he wrongly reports as Catuspithanibandha: 
samvat 100 60 5 srdvanasukladasamydm sukradine | rdjye sribhdskaradevasya 
| srigunakdmadevakdrite sripadmacakramahdvihdre sthitasdkyabhiksukumdra- 
candrena likhitam 'Copied by Sakyabhiksu Kumaracandra, resident of the Padma- 
cakramahavihara founded by Gunakamadeva, on Friday, the bright tenth of 
Sravana, in the year 165 during the reign of Bhaskaradeva'. The date of copying is 
26 July 1045 (Petech, loc. cit.). 


Genesis and Development of Tantrism 

(Dau Bahah) to Rudradeva I (c. 1007-1018) or Rudradeva II (1167-1175), 
the Hiranyavarnamahavihara (Kwa Bahah) and the Paravatamahavihara 
(Itum Bahah) to Bhaskaradeva (1039-1048), the Mayuravarnamahavihara 
(Bhiche Bahah) to Sankaradeva (1069-1082), the Tedovihara (Te Bahah) 
to Sivadeva (1098-1126), the Jayamanoharavarnamahavihara (Su Bahah) 
and Asanalokesvaramahavihara, also called Kacchapalagirimahavihara (Co 
Bahah) to Indradeva (1126-1136), the Cakravarnamahavihara (Cuka Bahah) 
to Manadeva (1136-1140), the Rudravarnamahavihara / Urikulimahavihara 
(Uku/U Bahah), the Manipurajaivamahavihara, and and the Bandhudatta- 
mahavihara to Narendradeva (1140-1147), and the Srivatsavihara (Atha 
Bahah) to Anandadeva ( 1147-1 167). 119 However, it is possible in the cases of 
Sankaradeva, Sivadeva, Manadeva, and Narendradeva, that the attribution 
intended was to their Licchavi namesakes. 

We have very little evidence for the reigns of these Thakuris, but what there 
is suffices to remove any suspicion that they were Buddhists to the exclusion of 
Saivism. According to the local chronicles Gunakamadeva made lavish donations 
to the temple of Pasupati, 120 Sankaradeva established a temple of a Siva with his 
name (Sankaresvara), 121 and Sivadeva gilded the roof of the temple of Pasupati, 

' For these monasteries and the names of the kings by whom they are said to have 
been founded (samskdrita- , kdrita-) see Locke 1980, pp. 32-33, and 1985, pp. 29, 
42, 74, 79, 82, 9L 95, 133, 140, 148. The dates of the reigns of these kings are as 
determined by Petech 1984. 

' Kaiser library Vamsdvali fragment (Petech 1984, Appendix), p. 2: raja sriguna- 
kdmadeva varsa 85 mdsa 6 || tena srlpasupatibhattdrakdya ekddasakosam pra- 
dattam tatraiva isdnesvarabhattdrakdya vdsukibhattdrakasya tdmmrasamsali- 
cchadanam krtya tatraiva *dlrgha*coparhika (conj. : copatrika Ed.) krtya tatraiva 
suvarnapanali kotihomam krtas ceti || raja, sri udayadeva varsa 6 || raja, srlnirbhaya- 
deva varsa 5 'King Gunakamadeva: 85 years and 6 months. He donated eleven 
[metal Lihga] sheaths to Lord Pasupati. At the same place he covered [the roofs of 
the shrines] of Lord Isanesvara and [the Naga] Lord Vasuki with copper *sheets 
(?), built a long rest-house and a golden water conduit, and performed a fire- 
sacrifice with ten million oblations'. King Udayadeva: 6 years; King Nirbhayadeva 
5 years || . . . '; cf. GopalarajavamsavalT f. 23vl-2: raja, srlgunakamadeva varsa 85 
ma 6 tena srlpasupatibhatarakaya ekadasa kosa sampradatta | tatraiva-m isanye- 
svarabhatarakaya tamrasamkhalasamchadanam krta | tatraiva dlrghacopa<r>hl 
krtah tatraiva suvarnapanali [kr]tah kotihoma purna krtam. The word samsali 
(=samkhali or samkhala) is evidently for Skt. srhkhald, srhkhalika 'chain'. I 
have conjectured the meaning 'sheet' considering the design of the Pasupati temple, 
whose roof is covered with interlocking metalic plates, panali = pranalika. With 
*copdrhi (conj.) cf. Classical Newari capdrha (Modern Newari capdh) 'rest-house' 
(Tamot et. al. 2000, s.v.). 

1 Kaiser library Vamsdvali fragment (Petech 1984, Appendix), p. 4: raja 
srisankaradeva varsa 17 | tena hi nandisdldydm samk<ar>resvarabhattd[rakd]ya 
pratisthitd devakulam ca purnam krtya rdstrasdntikd + + + + vihdras ca 
prdra<bha>ta 'King Sankaradeva: 17 years. He established [a Linga] for Lord 
Sankaresvara and completed a temple [for him]. He also undertook the con- 


The Saiva Age 

replaced the god's silver lotus, and donated a golden image of Siva. 122 Both In- 
dradeva and Anandadeva have the epithet paramasaiva- attached to their names 
in the colophons of manuscripts copied during their reigns; 123 and an inscription 
of 1143/4 records that Anandadeva, while he was the heir apparent (Yuvaraja), 
received Saiva initiation from the Saiddhantika Guru Rudrasiva of Benares, to- 
gether with the princes Vasantadeva, Somesvara, Yasomalla, and Arjunadeva: 124 

struction of the . . . monastery in order to avert danger from the kingdom' (I con- 
jecture rdstrasdntikdrandt for rdstrasdntikd + + ); cf. Gopalavamsavali f. 24rl— 
2: raja sri sahkaradeva varsa 15 tena ca namdisdldydm sahkaresvarabhatdraka 
pratisthitam tdmrasamchanna krtam devalam || puna bhagavati manahara 
bhatdrikd pratisthitd \\ rdstrasdnti bhavatih 'King Sahkaradeva: 15 years. He es- 
tablished [a Linga of] Sankaresvarabhattaraka at Nandisala and covered the tem- 
ple with a copper roof. He also established Bhagavati Manahara. This brought 
about the averting of danger from the kingdom'. 

122 Kaiser library Vamsdvali fragment (Petech 1984, Appendix III), pp. 4- 
5: raja, srisifvadeva vajrsa 27 mdsa 7 | tena hi pasupatibhattdrakasya 
suvarnasrm[khali]chddanam krta . . . srimatpasupatibhattdrakasya rajatapadma 
punar ghatita 'King Sivadeva: 27 years and 7 months. He covered [the temple 
of] Pasupatibhattaraka with gilded metal plates and remade his silver lotus'; cf. 
Gopalavamsavali f. 24r3-vl. 

123 Petech 1984, p. 57, colophon of a manuscript of the Cdndravydkaranavrtti in Ti- 
bet: srimadrdjddhirdjaparamesvaraparamabhattdrakaparamasaiva-indradeva- 
sya sri-indradevasya vijayardjye; and Petech 1984, p. 61, colophon of an Astasd- 
hasrikd Prajhdpdramitd manuscript: + + + paramabhattdrakaparamasaivama- 
hdrdjddhirdjasrTmaddnandadevapravarddhamdnakalydnavijayardjye. The scribal 
date of completion falls in 1134 in the first case and in 1166 in the second. 

124 Vv. 23-25: asydm sriraghuvamsamauktika*manir jdto jandnandanah 
(ACHARYA :mani . . .datah Regmi) sdndras candra ivdnvito 'timadhurair dnanda- 
devah karaih | uccaih saktidharah kumdrapadavim *prdpto 'pi tair (ACHARYA :prd- 
ptocitair Regmi) *diksito [ddntah siddhim avarna]niyamahimd (ACHARYA :diksita 
...ya mahimd Regmi) *prdpat pardm aisvarim (ACHARYA :prdpa . . . tyaisvarlm 
Regmi) || 24 *saurye 'rjunasamah (Achary A :saurye 'yam na sama Regmi) 
*preksya gundms tesu gunapriyah (ACHARYA : preksagunds te pragunapriyah 
Regmi) | bhaktim *arjunadevo 'pi vidadhe vibudhesv iva (ACHARYA : bhaktim arju- 
nam datvd ...vah) Regmi) 25 vasantadevo vijndni *dhimdn (ACHARY A :srimdn 
Regmi) somesvaras tathd | yasomalla*s ca (Acharya : sva Regmi) tair eva kumdrd 
diksitd ami. The plural pronouns here, tair diksito in 23c, gundms tesu in 24b, and 
tair eva in 25d, are plurals of respect (ddare bahuvacanam) and refer to Rudrasiva, 
who is also referred to in the plural in v. 12: sisyd babhuvur iha rudrasiva iti, as 
is his Guru Murtisiva in v. 8: bhattdrakd uditamurtisivdbhidhdndh. This record 
that contains these verses, a stone inscription now in the Government Museum in 
Kathmandu, has been published by Regmi (1965-1966, pt. 3, pp. 13-16) and, in a 
more complete and accurate form, by ACHARYA (1997) with an annotated Nepali 
translation. It was subsequently published by Tandan (1999, part 2, pp. 114-123), 
adopting only some of Acharya's improvements. ACHARYA understands the 
number 64 in the damaged penultimate line (. . . [cajtuhsasti . . . ydta sa .. . ) to be 
the last two digits of the inscription's date. The full number he conjectures to have 
been 264, which corresponds to A.D. 1143/4. He is surely right, since this is the 
only +64 date that fits the persons mentioned. Moreover, falling four years before 
Anandadeva became king the date accords with the information that he was still 


Genesis and Development of Tantrism 

In this [city] was born Anandadeva, a jewel in the pearl-necklace of the lineage 
of Raghu, delighting the people like a gentle moon with its most charming rays. 
Being self-controlled and of indescribable greatness, though he had achieved the 
status of prince (kumdra-) of great power (/though he had achieved the status 
[only] of Kumara who brandishes the javelin), he achieved when initiated by 
[Rudrasiva] the ultimate attainment of Sivafhood]. Likewise Arjunadeva, Ar- 
juna's equal in martial valour and a lover of virtues, conceived as great a de- 
votion to this [Rudrasiva] as to the gods, when he had seen his virtues. As for the 
learned Vasantadeva, the wise Somesvara, and Yasomalla, those princes too were 
initiated by the same [Guru]. 

Neither Arjunadeva nor Yasomalla are otherwise known from this ill- 
documented phase of Nepalese history. But we do have records of both a 
Vasantadeva, who was born in 1112 and died in 1163 but did not rule, and of 
a Somesvaradeva, who was born in 1119, died in 1182, and ruled from 1178 to 
1 183/5. 125 

The Bhauma-Karas ofOrissa 

But it was in the region of the modern territories of Bihar, West Bengal, 
Bangladesh, and Orissa that Buddhism enjoyed its most spectacular success in 
these centuries. It is only there that we find dynasties whose commitment to 
Buddhism was such that it was commonly signalled in their inscriptions through 
the use of such epithets as paramasaugatah and paramatathagatah 'entirely 
devoted to the Buddha'. Notable among these are the early Bhauma-Karas of 
Orissa (r.c. 825-950), 126 the early Candras of southeast Bengal (r. c. 850-1050), 
and, above all, the Pala emperors of Gauda (r.c. 750-1199), who at the height of 
their power extended their authority throughout eastern India and beyond. 127 

the Yuvaraja at the time of his initiation. 

125 See Petech 1984, pp. 64-67 and 71-72, and the Genealogical Table A, p. 229. 

126 rp^g name Bhauma-Kara is Indological. The early inscriptions speak of these rulers 
as Bhaumas and the later as Karas, evidently after the -kara that ends most of their 

127 The Palas and their successors, the Senas, are regularly described as kings of Gauda 
(gaudesvarah, gaudendrah, gaudardjah, gaudddhipah, gaudapatih, etc.); see, e.g., 
Sircar 1983a:26, 1. 33 (Laksmanasena); here pp. 108 (Nayapala) and 109 (Palapala, 
Mahlpala); Saduktikarnamrta 1449, 1496. The name Gauda in its narrow sense 
refers to a territory covering parts of West Bengal, being distinguished from Mag- 
adha, Variga, and Ariga. But with expansion of the power of its rulers it came to 
denote a much larger territory. Thus Campa in modern Bihar, the capital of ancient 
Ahga, is described as the capital of Gauda in the Anarghardghava (Act 7, prose be- 
fore v. 124: campa, nama gauddndm . . . rdjadhdnl), and Kausambi, about 35 miles 
south-west of Allahabad, is said to be in it in the Hitopadesa (Mitraldbha, Kathd 5, 
p. 19: asti gaudavisaye kausambi nama nagari). 


The Saiva Age 

Of the early Bhauma-Kara kings of Orissa Ksemarikara, who probably 
reigned around the beginning of the ninth century, is described in inscriptions as 
a paramopasakah 'a dedicated lay Buddhist', his son and successor Sivakara I as 
paramatathagatah, his son and successor Subhakara I, as paramasaugatah and 
paramopasakah, his son and successor Sivakara II as srisugatasrayah 'having 
the venerable Buddha as his refuge', and his son Subhakaradeva II, who reigned 
after his father's brother Santikara I alias Gayada, as paramasaugatah. 128 
A copper-plate of TribhuvanamahadevI, the Vaisnava (paramavaisnavi) wife 
of Santikara I, who occupied the throne as queen after the reign of her son 
Subhakara III alias Kusumahara, records that Subhakara (I), her husband's 
father, built a lofty Buddhist monastery; 129 another issued by her records 
that the earlier kings of her line had adorned the land with many Mathas, 
Buddhist monasteries, and temples; 130 and a third issued c. 980 by the para- 
mamahesvarah Sivakara III alias Lalitahara, the son of her grandson Sivakara 
II, records the granting of a village in favour of a temple of the Buddha in 
Uttaratosali made through him by his vassal Ranaka Vinitatunga. 131 

This epigraphical record is meagre, but it is very likely that it was the pa- 

128 EI 15:1 (the Neulpur grant of Subhakara I), 11.2-5, and EI 28:36 (the Terundia 
plate of Subhakara II), 11. 4-13. The religious affiliation of Santikara I and of five 
of the subsequent twelve rulers of this dynasty is not recorded. Among the remain- 
der are two Saiva kings, Subhakara IV and his brother and successor Sivakara 
III, two Vaisnava queens (paramavaisnavi), namely TribhuvanamahadevI I, wife of 
Santikara I, and TribhuvanamahadevI II, wife of Subhakara IV, and three Saiva 
queens (paramamahesvarl), Dandimahadevi, daughter of of Gaurimahadevi, wife 
and successor of Subhakara V, Vakulamahadevi, another wife of Subhakara V, and 
Dharmamahadevi, her successor and the wife of Santikara III. For the approximate 
dating of these rulers I follow D.C. Sircar's position (1953; EI 29:26, pp. 183-184 
and 189-191 [note 2]; Salomon 1998, pp. 190-191) that the Bhauma-Kara era be- 
gan c. 831. The Neulpur grant of Subhakara I was issued in year 8 of this era (EI 
15:1, 1. 30), i.e. c. 838, and the Terundia plate of Subhakara II in year 100 (EI 28:36, 
1. 22), i.e. c. 931. The last recorded date is 204 in the reign of Vakulamahadevi, i.e. 
c. 1035. 

129 EI 29:30, Baud plate A of TribhuvanamahadevI, 11. 5-6: sutottamas tasya 
samasrayafh] sriyah prasasad urvim susubhe subhakarah | kaler alahghyam 
sukrtasrayaya yo viharam uccair vidadhe silamayam 'His superlative son 
Subhakara, the resort of good fortune, [next] excelled ruling the land. To embody 
his merit he built a lofty monastery of stone which the degenerate age could not 

130 Shastri 1916:G, 11. 7-9: nirantaraviracitavividhamathaviharaprasadapraba- 
ndhaih purandarapurarohanasopanabandhair iva manditamahlmandalesv akha- 
ndalaprabhavesu maharajesu vyatltesu 'After the passing of those Maharajas, 
mighty as Indra, who adorned the land with the manifold sequences of Mathas, 
Viharas, and temples that they constructed without interruption as though with 
stairways for ascending to the heaven of Indra . . . '. 

131 Misra 1934:1, Talcher plate of Sivakaradeva, 11. 25-29. 


Genesis and Development of Tantrism 

tronage of these kings that enabled Mahayana Buddhism to grow and prosper as 
it did in Orissa, with the Tantric forms of that religion coming to the fore from 
the eighth century. 132 This efflorescence is attested by both archaeology and tex- 
tual evidence. Excavations of the Ratnagirimahavihara in the Cuttack district, 
not far from Guhesvarapataka, the Bhauma-Kara capital at or near the modern 
Jajpur, have revealed that this foundation underwent phenomenal expansion up 
to the twelfth century, 133 and this is only the foremost of several Buddhist sites 
in Orissa in which Tantric Buddhism is evident in the surviving statuary 134 The 
extremely high quality of Ratnagiri's stone-work renders it improbable that it 
was not a royal foundation. We have at least one Tantric text that reports that it 
was written here: the Samvarodaya nama Mandalopayika of Bhuvacarya, which 
survives in a Nepalese manuscript copied in 1050 in the Manadevamahavihara 
(Chuka Bahah); 135 and a manuscript of the Vimalaprabhd, the great commen- 
tary on the Kdlacakratantra, penned in the early decades of the twelfth century, 
in the thirty-ninth year of the reign of Harivarman, has a postscript in another 
hand added seven years later which locates the manuscript not far from Ratna- 
giri near the Benga river. 136 Indeed Ratnagiri had a particularly close associa- 
tion with the propagation of that Tantra according to the Tibetan account of the 

132 m itra 1981, pp. 20-21. Xuanzang reports in the early seventh century that Bud- 
dhism was the principal faith of the region, with some 100 monasteries and 10,000 
monks, all following the Mahayana; Xiyuji, p. 204. 

133 Mitra 1984, p. 225-232. On the phases of construction at Ratnagiri see Brown 
1978. On the successive phases of the Mantranaya manifest in the images that 
have survived at Ratnagiri and other Orissan sites see Linrothe 1999, pp. 53-57, 
70, 108-111, 125-128, 168-169, 195-198, 251-255, 280-283, and 287-288. 

134 Notable are the nearby sites of the Madhavapuravihara at Udayagiri and and the 
Candradityavihara at Lalitagiri. On Udayagiri see Bandyopadhyaya 2007; and 
on Lalitagiri see Chauley 2000; and IAR 1985-6, pp. 62-63; 1986-87, pp. 64-67; 
1987-88, pp. 88-90; 1988-89, pp. 65-66; 1989-90, pp. 77-80; 1990-91, pp. 54-55. 

135 Samvarodaya f. 56v3-4: srlmadratnagirau sthitva sarvasattvarthahetuna | krte- 
yam mandalopayika bhuvaca<r>yena dhimata | srlsamvarodaya nama mandalo- 
payika *samapta (corr. : samaptah Cod.) || • || samvat a cu *prosthapadakrsnaca- 
turthyam (prosthapada conj. : pretipada Cod.) rajadhirajapa[ram]esvaraparama- 
bhattarakasribaladeva + + vijayaraj<y>e likhitam \ srimanadevamaha*vihariyasa- 
kyabhiksusadhusridevasya (vihariya conj. : vihare Cod.) pustakam *<| yad atra 
punyam tad bhavatu> (diagn. conj.) matapitrgurupadhyayasakalasattvarase<r> 
anuttara<jnana>phala*praptaya iti (conj. : prapnoti Cod.). 

136 Shastri 1917, pp. 79-80 (ASB MS 10766). The manuscript is dated by the scribe 
in year 39 of the reign of Maharajadhiraja Harivarman, on whom see Majumdar 
1971, pp. 209-210. Colophon: maharajadhirajasrimat-harivarmadevapadiyasam- 
vat 39 | suryagatya asadhadine 39. The postscript: satcatvarimsatigate vatsare 
harivarmanah \ maghasya krsnasaptamyam ekadasadine gate || mrtaya cuncaduka- 
ya gauryd svapnena drstaya | kanisthahgulim adaya *prstayedam (corr. : prstha- 
yedam Shastri) udlritam | purvottare disobhage bemganadyas tatha kule | \pacca- 
tvam bhasitavatah] saptasamvatsarair iti. 


The Saiva Age 

history of the transmission of its teachings maintained in the lineage that de- 
scends from Rva chos rab in the early twelfth century For that relates that the 
Vimalaprabhd was transmitted by an emanation of Manjusri to Pandita Cilu, a 
native of Orissa trained at the Ratnagiri monastery and reached Rva chos rab 
after being passed on through five intermediaries in Bengal and Bihar. 137 A tra- 
dition that Cilu studied the Kdlacakratantra in the Ratnagirimahavihara before 
seeking the Vimalaprabhd is recorded by Gzhon nu dpal. 138 

The Candras of South-East Bengal 

As for the Candras, they used the wheel of the Buddha's teaching (dharma- 
cakram) as the seal-symbol on their charters; the Pascimbhag copper-plate grant 
of Sricandra I (r. c. 925-75) describes both this king and his predecessor Trailoky- 
acandra as paramasaugatah; 139 and his Rampal and Madanpur copper-plate 
grants describe Suvarnacandra, the predecessor of Trailokyacandra (r. c. 900- 
925), as a bauddhah 'a follower of the Buddha's teachings'. 140 After Trailokya- 
candra came Sricandra (II), Kalyanacandra, Ladahacandra, and Govindacandra. 
The Mainamati plates of Ladahacandra and Govindacandra (r. c. 1000-1020 and 
c. 1020-1045) provide these names and reveal that the last two were parama- 
saugatah. 141 

The Khadgas of Samatata 

We have epigraphical evidence of three successive generations of kings of the 
Khadga line ruling the Samatata region of southeast Bengal from about 625 into 

137 OROFINO 1994, pp. 17-23; Blue Annals, p. 755. 

138 Blue Annals, p. 755. 

139 EI 37:51, 11.25-26. 

140 EI 12:18, 1. 6; EI 28:9, 1. 8; and Majumdar 1971, p. 201. 

141 EI 38:35, no. 1, 11. 35-36; no. 2, 11. 6-7; no. 3, 11. 33-34. As for Purnacandra (r. c. 850- 
875), there is no explicit evidence of his religious persuasion. MAJUMDAR (1971, 
p. 201) argues that since it is said in the Rampal copper-plate that Suvarnacandra, 
his son, "became a follower of the Buddha" (EI 12:18, 11.5-7) it is probable that 
before him the family was non-Buddhist. This is not accurate, since the text says 
not that he became a Buddhist but only that he was one: buddhasya yah sasaka- 
jatakam ahkasamstham bhaktya bibharti || bhagavan amrtakaramsuh | candrasya 
tasya kulajata itiva bauddhafh] putrah sruto jagati tasya suvarnacandra!} 'His son 
was Suvarnacandra, famed in the world, a Buddhist as though [simply] because he 
was born in the lineage of the Moon (/the Candra lineage), which out of devotion to 
the Buddha displays his incarnation as a hare in its markings'. The allusion here 
is to the story exemplifying the Buddhist Perfection of Generosity (danaparamita) 
that the Buddha gave away his own body as food when he was a hare in a former 
life, the sasajatakam. The immediately preceding verse, which is devoted to Purna- 
candra, says nothing substantive about him but only that his name is found as that 
of the first of the kings of this dynasty in Prasastis and other inscriptions. 


Genesis and Development of Tantrism 

the early years of the eighth century. Though the inscriptions do not include the 
epithet paramasaugatah they do speak of these rulers in equivalent terms. The 
first, Khadgodyama, is described in an inscription of his great-grandson Rajaraja 
as having conquered the earth after declaring his intense devotion to the Three 
Jewels: the Buddha, his teachings, and the Sahgha. 142 The same inscription tells 
us that Rajaraja gave land to these three; 143 and another that Devakhadga, the 
father of Rajaraja, made a donation to the same for the longevity of his son. 144 
We have no evidence of any support given to Saivism by these kings themselves. 
But a pedestal inscription on an image of the Saiva Goddess records that it was 
gilded out of devotion by Prabhavati, Devakhadga's queen. 145 

The Candras ofArakan and Miscellaneous Other Buddhist Kings of Eastern In- 

That there were Buddhists among the Candras ofArakan is evident from the 
Mrohaung pillar inscription of Anandacandra, which has been dated around the 
end of the third decade of the eighth century 146 This gives a list of the names 
and reign-durations of the kings who preceded him from c. 380 onwards with 
an interruption of unspecified length. After this interruption come the rulers 
of the Candra dynasty down to Anandacandra himself, spanning in this second 

142 Ashrafpur plate B (Laskar 1907), 11. 2-4: trailokyakhydtaklrtau bhagavati sugate 
sarvalokfe] + + + taddharme sdntarupe bhavavibhavabhiddm yogindm yoga*gamye 
(corr. : gamya Ed.) | tatsahghe cdprameye vividhagunanidhau bhaktim dvedya 
gurvlm srlmatkhadgodyamena ksitir iyam abhito nirjitd yena 'Khadgodyama, who 
conquered this earth in all directions after declaring his intense devotion to the 
Lord Buddha, whose glory has been declared throughout the three worlds, among 
all men . . . , to his tranquil teachings that can be realized by Yogins who [therebyl 
break the power of [transmigratory] existence, and to his numberless Sangha, the 
repository of manifold virtues'. 

143 Ashrafpur plate B (Laskar 1907), 11. 6-7: tatsuto rdjardjah dattam ratna- 
traydya tribhavabhaya*bhide (conj. : bhidd Ed.) yena ddnam svabhumeh 'His 
[Devakhadga's] son, who made a gift of his land to the Three Jewels that elimi- 
nate the fear of the three worlds'. To give to the Three Jewels is, I surmise, to make 
a grant to be divided between the Buddha for the building or maintenance of Bud- 
dhist shrines (gandhakuti) and Stupas, the Dharma for the copying and teaching of 
sacred texts, and to the Sangha for its sustenance and comfort. 

144 Ashrafpur plate A (Laskar 1907). 

145 EI 17:24,4, 11. 1-2: taddtmajo ddnapatih pratdpT srldevakhadgo vijitdrikhadgah 
| rdjhas tasya mahddevT mahisi srlprabhdvati \ sarvdnipratimdm bhaktyd 
hemaliptdm akdrayat 'His son was the majestic donor (ddnapatih) Devakhadga, 
whose sword had defeated his enemies. The chief consort of that king, Mahadevi 
Prabhavavati, had [this] image of Sarvani gilded'. The word ddnapatih is the stan- 
dard Buddhist term for one who gives to monks, the Dharma, or the Buddha. The 
image (Huntington 1984, fig. 26) was found in the village of Deulbadi, near 
Comilla, together with a Surya and small Lirigas, all of brass. 

146 D.C. Sircar in EI 32:11, p. 1071-108. 


The Saiva Age 

sequence a total of three hundred and fifty years. For most of his ancestors we 
are given no information other than their names and the lengths of their reigns, 
but the record is more forthcoming as it approaches the time of Anandacandra 
himself. Vajrasakti (r. c. 649-665) is said to have died and gone to the world 
of the gods endowed with [the Buddhist perfections (paramitah) of] generosity, 
morality and the rest, and his successor Dharmavijaya (665-701) is said to gone 
to the same, this time defined as the Buddhist Tusita heaven, as a result of his 
firm commitment to the Three Jewels. 147 Two short inscriptions from Vesali of 
the time of his ancestors Niticandra (r. c. 520-575) and Viracandra (r. c. 575- 
578) tell us that the wife of the former, queen Savitam-Candrasri, was a lay 
Buddhist (paramopdsikd) and that the latter established a hundred Stupas. 148 
As for Anandacandra, he calls himself a lay Buddhist and devotes nine verses to 
detailing his works of Buddhist piety, which included building many monasteries 
with his own name, establishing precious images of Buddhas, Bodhisattvas, and 
such [Mahayanist] goddesses as Cunda, having hundreds of Buddhist scriptures 
copied, and giving to many monks from various lands, which is to say, that he 
fulfilled to the best of his ability his duty to honour each of the Three Jewels. 149 
Yet even this devoted patron of his faith did not neglect to extend his support 
to the followers of other religions in his realm. He tells us that although he is a 
Buddhist he desires the good of all beings, lest his cultivation of the Buddhist Per- 
fection of Generosity (danaparamita) be incomplete, and so has established four 
Mathas for the housing of fifty brahmins, providing them with land and workers, 
and two others, the Anandesvaramatha and the Anandamadhavamatha, whose 
names reveal that they were associated with a Siva and a Visnu established with 
his name. 150 Moreover, a fragmentary copper-plate inscription (EI 37:13) from a 

147 Inscription of the western face of the pillar at the Shittaung Pagoda, Mrohaung, 
Arakan (JOHNSTON 1944:A), vv. 37c-40: vajrasaktis tata<h> [khjyato raja deva- 
nvayodbhavah || pratipalya jagat sarvam rajyam sodasavatsaram | danasiladi- 
samyukto devalokam. sa yatavan \\ sridharmajayasamyukto lokanugrahatat- 
parah | tatpascad abhavad dhirah sridharmavijayo nrpah || sattrimsad abdany 
upabhujya rajyam dharmena nitya ca jayena caiva \ ratnatrayanusmaranabhi- 
yogat sa devalokam tusitam prayatah . 

148 EI 32:11, no. 1, 11.3-4: devisavitam-candrasrlya nama paremopasikasya; EI 32:11, 
no. 2, 11. 1, 3—4: satyadharmmanaragena krtam svarthena bhubhuja . . . srivira- 
candradevena mahimandalamandanam | dharmmadhigatarajyena buddhastupa- 
satam krtam. 

149 Johnston 1944:A, vv. 46-54. 

150 Johnston 1944:A, vv. 55-56: pahcasadbrahmanavasam ksetrabhrtyasamanvitam 
| vadyavadakasamyuktam karitam mathacatustayam || 
mathas canandamadhavah \ anandesvaranamapi naulakk[e] ca matha<h> smrtah. 
The practice of establishing a Visnu with the founder's name followed by -madhava 
(as an alternative to the standard -svamin) is in accordance with textual pre- 
scription; see Somasambhu, Brunner 1998, p. 311 (v. 48), =Kriyakandakramavall, 


Genesis and Development of Tantrism 

site near Mrohaung recording a donation by queen Kimmajuvdevi of a village to 
a Buddhist monastery founded by herself begins by relating six generations of 
the ascendants of her husband the king. Unfortunately the names of this king 
and his ancestors have been lost through the scissoring off of strips from the top 
and right hand side of the plate. However, what remains conveys the unexpected 
information that all these kings were paramamahesvarah. The editor of the in- 
scription assigns it to the sixth century on the grounds of its close palaeographic 
similarity to the grants of Niticandra and Vlracandra, and argues that if the first 
of the six kings was, as is likely, Dvehcandra, the founder of the Candra dynasty, 
then the king in question was Niticandra's father Bhuticandra (r. c. 496-520). 151 
Viracandra, he argues, is excluded by the fact that one of the two Vesali inscrip- 
tions records his patronage of Buddhism. However, that a king should give to 
Buddhism and at the same time be declared a paramamahesvarah in documents 
issued by the royal chancellery is quite within the bounds of possibility, as we 
have seen. 

Other royals of eastern India who are identified as paramasaugatah in our 
period — apart from the imperial Palas, to whom I shall turn presently — are 
Bhavadeva of Devaparvata in Samatata (r. c. 765-780), the founder of the 
Buddhist monastery Bhavadevamahavihara at Pattikera, modern Mainamati, 
Rajyapala of the Kamboja dynasty of Priyahgupura in the tenth, Madhusena, 
the Sena king of Gauda, in the thirteenth, and, in Orissa, Udayavaraha of 
the Mayuravamsa at some time in the tenth to twelfth, the Nandodbhava 
Dhruvananda of Jayapura, the successor of the paramamahesvarah Devananda 
II, in the late tenth, and Kantideva of Harikela in the ninth. 152 The inscription 
that tells us that the last was paramasaugatah also conveys that Buddhism was 
the faith of his grandfather Bhadradatta. After a benedictory verse in praise of 
the Buddha it begins the eulogy of the donor's forebears with this king, saying 
that his devotion to the Buddha had intensified his power and that he had 
[thereby] conquered all his enemies. His son Dhanadatta, the donor's father, is 

ff. 72v7— 73rl: svamyantam madhavantam va kartrnamnd ca samyutam | dharayen 
ndma devasya visnoh sthapanam iritam 'He should bestow a name on the deity 
conjoined with the name of the patron and ending in -svamin or -madhava. I have 
[thusl explained the installation of Visnus'. 

151 D.C. Sircar, EI 37:13, p. 64. 

152 Sircar 1983a, Supplements, 11. 42-43 (Bhavadeva); Mitra 1971, p. 245 (Bhavade- 
vamahavihara). EI 41:22, 11. 19-20 (Rajyapala); the final colophon of ASB, MS 
40785 dated in 1289; see Shastri 1917, p. 117 (Madhusena). Shastri 1920, p. 243, 
11. 2-3, 6 (Udayavaraha). Tripathy 1930, p. 466, 1. 24 (Dhruvananda). EI 29:26, 
11. 25-26 (Devananda). EI 26:45, 1. 14 (Kantideva). The exact location of Harikela 
is uncertain, but it may be placed with some confidence in the area of Chittagong, 
that is to say, near Samatata in the direction of Arakan. 


The Saiva Age 

praised only for his learning in poetry, the Epics, and the Puranas. Mention is 
made not of his religion but of that of his wife Bindurati, who is said to have 
been a devotee of Siva. 153 

The Pala Emperors and the Great Monasteries of Eastern India 

With the Pala emperors we come to what appears to be the most robustly 
Buddhist of all the dynasties of our period. Like the Candras of southeast Bengal 
they chose the wheel of the Buddha's teaching (dharmacakram) as the seal- 
symbol on their charters; they began their inscriptions with obseisance to the 
Buddha; and the following among them appear with the epithet paramasaugatah 
in the lacunose record of inscriptions and manuscript colophons: Dharmapala 
(r.c. 775-812), Devapala (r.c. 812-850), Mahendrapala (850-865+), Narayana- 
pala (r.c. 865+-917), Vigrahapala II (r. c. 972-977), Mahipala I (r. c. 977-1027), 
Nayapala (r. c. 1027-1043), Vigrahapala III (r.c. 1043-1070), Ramapala (r. 
c. 1072-1126), and Madanapala (r. c. 1143-1161). 154 

Under these rulers eastern India witnessed an extraordinary development 

' EI 26:45, 11. 3-: . . .jayaty udaro durvaramaravisarasya jayl jinendrah || tad- 
bhaktibalitasaktir bhujadvayaurjityavijitaripudarpah | sa jayati dharmaikaratah 
khyatah sribhadradatto yah || tasya subhasitabharatapurdnaramdyanarthavit 
tanayah \ namna sridhanadattah prakatitamahimanvayo yo 'bhut || tasya gauri 
mahabhubhrtsuta budhagurustutd \ patni binduratir ndma yd babhuva sivapriyd 
'Victorious is the foremost of the Jinas, the exalted one who conquered the multi- 
tude of Maras so hard to ward off. His power intensified by devotion to him, the 
pride of his enemies overcome by the strength of his two arms, solely devoted to 
the Dharma, victorious is the famous Bhadradatta. His son was Dhanadatta. He 
understood the meaning of elegant poetry, the Mahabharata, the Puranas, and the 
Ramayana, and his uninterrupted greatness was made manifest [to all]. His wife 
was Bindumati, the fair-skinned daughter of a great king, praised by the learned 
and her elders, a devotee of Siva'. 

Dharmapala: EI 4:34, 11. 29-30; EI 17:17, 11. 24-25; EI 18:30, 1. 28. Devapala: EI 
17:17, 11. 24-25; EI 18:30, 1. 29. Mahendrapala: EI 42:2, 11. 30-31. Narayanapala: 
Sircar 1983a:17, 11.28-29. Vigrahapala II: EI 29:1A, 11.27-28. Mahipala I: EI 
14:23, 11.29-30; EI 29:1, 1.27; a pedestal inscription (HUNTINGTON 1984, pp. 221- 
222). Nayapala: colophon of a MS transcribed in Bendall 1883, p. 175. Vigra- 
hapala III: EI 15:18, 1.23; EI 29:1B, 11.26-27; EI 29:7, 11.24-25; MS colophon 
transcribed in Bendall 1902, pp. 232-233 (because the date of copying is said 
here to be the 26th year of the reign of Vigrahapala this can only refer to Vigra- 
hapala III). Ramapala: Regmi 1965-1966, Pt. 1, p. 148 (MS colophon); colophon 
oiKubjikamata, NAKMS 1-1633, NGMPP B25/22 (transcribed in GOUDRIAAN and 
Schoterman 1988, p. 6); a pedestal inscription (Huntington 1984, pp. 233-234). 
Madanapala: Mukherji and Maity 1967:30, 11.31-32. The dates of the reigns 
given here are those proposed by D.C. Sircar (1975-1976), with the addition of 
those of Mahendrapala. The existence of a Pala Mahendrapala, son and successor 
of Devapala, was established only with the publication of the Malda District Mu- 
seum copper-plate charter of that king in 1992 (EI 42:2) by K.V. Ramesh and S. 
SUBRAMONIA Iyer, following its discovery in 1989. 


Genesis and Development of Tantrism 

of Mahayana Buddhism in all its branches, particularly in the Tantric Way of 
Mantras (Mantranaya), 155 which if not entirely the product of this region was 
very largely so; and this immense creativity, whose products formed in due course 
the basis of the Buddhism of Inner Asia, was nurtured and refined in a num- 
ber of major monasteries, of which the most eminent were those of Nalanda, 
Vikramaslla, Somapura, Trikatuka, Uddandapura, and Jagaddala. 156 That the 

155 rp^g rj er g e edition of the Tripitaka contains 486 works (Toh. 360-845) in the 
section of the Kanjur devoted to scriptural Tantric works and 2606 (Toh. 1180- 
3785) in the section of the Tenjur devoted to works of Tantric scholarship, com- 
prising commentaries on the Buddhist Tantras and works setting out observances 
(Sadhana, Bali, Pratistha etc.) based on them. All claim to be translations of 
Sanskrit originals and this claim is true in the great majority of cases. In ad- 
dition there are numerous works surviving whole or in citation in Sanskrit that 
appear not to have been translated into Tibetan; and some of these, such as 
the Gudhapadd of Advayavajra, the Mandalopdyikd of Padmasrimitra, the Va- 
jrajvdlodayd of Anandagarbha, the VajravdrdhTkalpa, the Sarvadevasamdgama, 
and the Herukasddhana of Kalyanagarbha, have been used in this study. 

156 rp^ e Nalandamahavihara was located in Bihar about 55 miles southeast of Patna, 
with the Uddandapuramahavihara close by. The Vikramasilamahavihara was very 
probably at Antichak in the Bhagalpur District of Bihar about 19 miles from 
Bhagalpur town. No evidence conclusively etablishes this. But the huge size of the 
monastery excavated at Antichak severely narrows the field of known possibilities; 
and there is suggestive archaeological evidence: a copper seal was uncovered in the 
ruins of the monastery with the legend vikramasya (IAR, 1973-4, p. 9) and a dam- 
aged inscription on a Stupa there contains the syllables vikrama. . . (HUNTINGTON 
1984, pp. 125-126). The use of Vikrama for Vikramaslla is seen in Anupamavajra's 
Adikarmapradipa; see here p. 91. That the name of the monastery was Vikrama- 
slla rather than Vikramasila, as it appears in some secondary sources, is clear from, 
e.g., the scribal colophon of a manuscript of Vajragarbha's Hevajratantrapindartha- 
tika that was penned there: srimadvikramasilamahdvihdre lekhapitam. The Soma- 
puramahavihara was at Paharpur about 29 miles northwest of Mahasthan (ancient 
Pundravardhana) in Varendri, the region of northern Bengal between the arms of 
the Ganges and Karatoya rivers (Rdmacarita 3.10ab: apy abhito gangdkaratoyd- 
narghapravdhapunyatamdm. The Jagaddalamahavihara too was in this region; 
see Rdmacarita 3.7: . . . jagaddalamahdvihdracitardgdm | dadhatlm lokesam api 
mahattdrodiritorumahimdnam '[the land (of Varendri)!, whose beauty was height- 
ened by the Jagaddalamahavihara, which was home to Lokesvara, its extensive 
glory proclaimed by [a] great [image of] Tara'. Its site has beeen tentatively iden- 
tified as the mound at modern Jagdal in the Dhamoirhat Upazila of the Naogaon 
District of the Rajshahi Division of Bangladesh. A one-season, small-scale exca- 
vation of this mound was undertaken by Bangladesh's Department of Archaeology 
in the winter of 1996. Though it revealed evidence of the presence of a Buddhist 
monastery and unearthed a fine statue of Heruka and his consort, most of the site 
was left untouched and nothing has been reported that raises to certainty the high 
probability that this was the Jagaddalamahavihara. See Zakaria 1994 and Miah 
1997/8. The location of the Trikatukavihara is as yet unknown, but Taranatha re- 
lates a myth that on instructions from Mahakala king Devapala unearthed this 
monastery beneath a sand dune when he was crossing Rara (=Radha) (HBI, p. 267; 
Majumdar 1971, p. 525), the region of Bengal south of Varendri and west of the 
Bhagirathi river, divided into Uttararadha, covering part of Birbhum District and 

The Saiva Age 

Palas' devotion to the Buddha was expressed, as we might expect, in the creation 
and support of these great monastic universities is shown by terracotta seals 
found amid their remains, and by the Rgya gar chos 'byung ('The Arising of the 
Dharma in India'), a Tibetan account of the history of Indian Buddhism written 
in 1608. 

Taranatha, the author of this work, tells us that he wrote it on the basis 
of three Sanskrit sources that are now lost or inaccessible. The first is an un- 
named work in 2,000 verses by a scholar of Magadha named Sa dbang bzang 
po, that is to say, Ksmendrabhadra or a synonym such as Dharanindrabhadra. 
This covered the history of the religion up to the time of the Pala king Ramapala 
(r. c. 1072-1126). The second is the Buddhapurana, a work by Dbang pos sbyin 
(Indradatta) in 1,200 verses, which went beyond Ramapala to cover the succes- 
sor dynasty of the Senas of Gauda. It may therefore be supposed to have been 
composed in that part of India, like the work of Ksmendrabhadra. The third 
is a work of similar length covering the succession of Acaryas and written by a 
brahmin scholar whom Taranatha calls Bhataghati. This name is implausible 
as it stands. If, as is probable, it is is deformation of Vandyaghatiya, then it 
identifies him as a member of a well-known Radhiya brahmin lineage of Ben- 
gal (> Bandyopadhyaya, Banerjee). 157 Taranatha claims to have relied primar- 
ily on the first of these three works, that is to say, for his account up to the 
time of Ramapala, since that source went no further. 158 For the period of the 
Senas, who succeeded the Palas, he must have relied on Indradatta alone. As for 
Vandyaghatiya's account of the succession of Acaryas, it is probable that it con- 
sisted of, or extended to, an account of the succession of the Tantric Acaryas of 
Vikramasila from its foundation in the eighth century to its destruction around 
1200 by the forces of Muhammad Bakhtyar Khalji. For he adds a section in the 
manner of a supplement on the Acaryas of Vikramasila after his treatment of 
the periods covered by his first two sources. His work, then, derives from Indian 
tradition, and while his sources were evidently inaccurate for the early history 
of Buddhism, we might expect them, particularly the work of Ksmendrabhadra, 

the whole of Burdwan District, and Daksinaradha, covering Bankura District and 
the non-coastal part of Midnapur District. 

157 In the eulogy of Bhatta Bhavadeva, the learned minister of Harivarman (c. 1090+), 
in a stone inscription from Bhubaneswar, Bhavadeva's mother Sangoka is said to be 
the daughter of a Vandyaghatiya brahmin (EI 6:17B, v. 13). Other Vandyaghatiyas 
are the Sarvananda who in 1159 wrote a commentary Tikasarvasva on the 
Lihganusasana of Amarasimha, the great 16th-century Dharmasastrin Raghunan- 
dana, author of the Smrtitattva (PlNGREE 1994, p. 341), Narayana (ft. c. 1681), 
author of the Smrtisarvasva or Smrtitattva (PlNGREE 1994, p. 181), and Dvija 
Laksmana, who translated the Adikanda of the Adhyatmaramayana into Bengali. 

158 Rgya gar chos 'byung, pp. 215, 1. 22-214, 1.10; HBI, p. 350. 


Genesis and Development of Tantrism 

to be more reliable in their account of what for them was recent history. 159 The 
Rgya gar chos 'byung therefore deserves close attention. 

Taranatha attributes to Dharmapala the building of the monastery of 
Vikramasila and to Devapala the building of the monasteries of Somapura and 
Trikatuka. 160 In this, however, he or his sources are confused. The claim that the 
monastery at Somapura was founded by Devapala is contradicted by a terracotta 
seal found at the site bearing the legend srisomapure sridharmmapdladeva- 
mahdvihdre 'in the Mahavihara of Dharmapaladeva at Somapura', 161 thereby 
indicating that it was founded not by Devapala but by his father Dharmapala. 
Evidence also contradicts Taranatha's claim that it was Devapala that built the 
Trikatuka monastery. For Haribhadra reports at the end of his Abhisamayd- 
lamkdrdloka, his great commentary on the Astasdhasrikd Prajndpdramitd, that 
he composed it in this monastery during the reign of Dharmapala and under his 
patronage. 162 

159 After his account of the Tantric Acaryas who held office successively at Vikramasila 
Taranatha offers brief treatments of various topics not covered by these sources. 
Buddhism in mainland Southeast Asia and in maritime Southeast Asia, Sri Lanka 
and the South is covered in ch. 39 and 40 respectively. On these topics, he says, 
he has seen no comprehensive work. Ch. 41 treats the spread of Buddhism in the 
Deccan following another lost work, the Flower-Garland, by a brahmin Manomati, 
which, he says, contained a brief account of this subject. Ch. 42 covers the divisions 
of the main Nikayas, evidently on the basis of such Indian treatments of the topic 
as the Samayabhedoparacanacakra of Vinitadeva; ch. 43 examines what he rightly 
considers to the muddled theories of the origin of the Mantranaya; and ch. 44 gives 
some notes on the various Indian schools of image-makers. This is followed by the 
account of his use of his sources. He notes that he has no written sources for the 
later events in his account that were not covered in those works. For these events 
he has relied on what he judged to be trustworthy oral reports. 

160 See Rgya gar chos 'byung, p. 160, 11. 9-10 (Somapuravihara); p. 161, 1. 11 (dpal tsha 
ba gsum gtsug lag khang [Trikatukavihara]; cf. p. 167, 11. 7-8: tri ka *tu [corr. : ta 
Ed.l ka tsha ba gsum kyi gtsug lag khang); p. 165, 1. 17 (Vikramasilavihara); HBI, 
p. 266, p. 267, pp. 274-275. 

161 ARE 1927-28, pp. 105-106; Dikshit 1938, pp.20 and 90, and plate UXh; N.G. 
Majumdar in EI 21:16, p. 98. 

162 Abhisamayalamkaraloka , p. 994, vv. 6— 7: khyato yo bhuvi punyakirtinicayo vidvaj- 
janalamkrtas tasmin sarvagunakare trikatukasrimadvihare subhe | danal labdha- 
mahodayasya karunadevasya dharmatmanah sanathyena sukhopadhananilaye 
sthitva vivekaspade || krudhyatkunjarakumbhapithadalanavydsaktasaktyatmanah 
punyabhyasakrtabhiyogajavasat sampatsamadayinah | rajye rajyabhatadivamsa- 
patitasrldharmapalasya vai tattvalokavidhayini viracita satpanjikeyam maya 'I 
have composed this excellent commentary that illuminates reality after taking up 
residence in the splendid Trikatukavihara that is famed throughout the world, the 
site of a mass of sacred edifices, adorned by learned men, a store of all the virtues, 
where [all] the means of happiness are to be found, a place of insight, through the 
support of the compassionate king Dharma[pala], who by means of donation has 
achieved pre-eminence[; and I have done so] during the reign of this king, who born 
in the dynasty that descends from Rajyabhata, full of power devoted to the rending 


The Saiva Age 

In the case of the Somapura monastery it has been argued that we may 
salvage Taranatha's credibility by concluding that Devapala did found this 
monastery, as Taranatha claims, and that he gave it his father's name rather 
than his own out of filial piety 163 This is indeed a practice of which there are 
other examples, its purpose being to transfer to the person named the religious 
merit generated by the creation and use of the foundation; but it is much more 
probable that Taranatha is in error here, as he clearly is in the case of the 
Trikatuka monastery For his history commits the fundamental error of revers- 
ing the true sequence of the two reigns, placing that of Devapala before that of 
Dharmapala. 164 His attribution of the founding of Somapura and Trikatuka to 
Devapala rather than Dharmapala can, then, readily be explained as the result 
of this reversal. We may therefore suspect that his attribution of the founding 
of Vikramasila to Dharmapala suffers from the same dislocation and that its 
true founder was his son Devapala. That this suspicion is correct is confirmed 
by the Adikarmapradipa of Anupamavajra. For in its conclusion he tells us 
that he compiled the work following the instruction of Dharmakara, a monk 
whom he describes as "residing in the monastery called Vikrama constructed 
by king Devapala". 165 Vikrama here is evidently a bhimavat contraction for 
Vikramasila. 166 However, we may not conclude that everything that Taranatha 
attributes to Dharmapala was Devapala's doing, and vice versa. He reports, for 

of the swollen globes on the foreheads of the furious elephants [of his enemies], has 
attained his glorious success by virtue of the dedication produced by his repeated 
pious works'. For the use of sthitvd here cf. the final verse of the Samvarodaya 
nama mandalopayika of Bhuvacarya cited here, p. 82. 

163 N.G. Majumdar in EI 21:16, p. 98, fn. 5. 

164 Rgya gar chos 'byung, chapters 29 (Devapala) and 30 (Dharmapala). Taranatha 
gives the order Gopala > Devapala > Rasapala > Dharmapala; see Rgya gar chos 
'byung, pp. 163-164: rgyal po de wa pa las lo bzhi bcu brgyad du rgyal srid byas | 
de'i rjes su sras ra sa pa la rgyal srid lo bcu gnis byas 'King Devapala ruled for forty- 
eight years. After him his son Rasapala ruled for twelve'. No Rasapala appears in 
the accounts of the dynasty given in the Palas' inscriptions. The name is perhaps a 
deformation of Rajyapala (r. c. 917-952), the successor of Narayanapala. 

165 Adikarmapradipa, ed. Takahashi, p. 153: vihdre (T [metri causa] : vihare P, Ed.) 
*nrpadevapalaracite (T, Ed. : ndapadevaracita P) *srlvikramakhye (T, Ed. : srlvi- 
kramaksa P) sthitah srimatsaugatasasanaikatilakah khyato 'dvitiyah krti | *sila- 
dhyas cirabrahmacaryacarito (P : siladhyasthiratattvadrstimahito T, Ed.) dharma- 
kara!} *santadhls (P : sanmatih T, Ed.) Hasyddesakarah samasty anupamah (T, Ed. 
: Hasyadesakaro babhuva 'nupamas P) tenadikarmoddhrtam '[This text on] the 
initial observance has been extracted [from various sources] by Anupama, acting on 
the instruction of Dharmakara, that renowned, unequalled scholar, richly endowed 
with morality, of tranquil mind, a life-long observer of celibacy, a resident of the 
Vikrama monastery constructed by King Devapala'. 

166 On Vikrama for Vikramasila see here p. 88. 


Genesis and Development of Tantrism 

example, Dharmapala's particular reverence for Haribhadra, 167 a relationship 
that, as we have seen, Haribhadra himself attests. He also claims that Dharma- 
pala created about fifty religious foundations {dharmadhikarah), and that the 
majority, thirty -five, were for the study of the Prajnaparamita texts. 168 It is 
at least probable that this bias was due to the influence of Haribhadra, given 
the latter's close relationship with Dharmapala and the fact that he was the 
pre-eminent scholar of his age in the exegesis of this literature. 

As for the monastery of Uddandapura, which was located near the more 
ancient monastery of Nalanda, Bu ston, in his history of Buddhism in India and 
Tibet, completed in 1322, attributes its foundation to Dharmapala; 169 and the 

167 Rgya gar chos 'byung, p. 167, 11. 7-9: de nas mi ring bar rgyal po dha rma pa las 
spyan drangs ste | tri ka *tu (corr. : ta Ed.) ka tsha ba gsum kyi gtsug lag khang 
du bzhugs nas | sher phyin nyan pa stong phrag mangpo la chos ston cing | brgyad 
stong grel chen la sogs pa bstan bcos kyang mang du mdzad 'Not long after this 
[Haribhadra! was invited by King Dharmapala. He stayed in the Trikatukavihara 
and taught the Prajnaparamita to many thousands of hearers. He also composed 
[his] detailed commentary on the Astasahasrika, and many other learned works'; 
HBI, p. 277. 

168 Rgya gar chos 'byung, p. 165, 11. 14—17: rgyal srid du 'khod ma thag nas shes rob kyi 
pha rol tu phyin pa 'chad pa po mams spyan drangs | slob dpon seng ge bzang po 
la khyad par du mos | rgyal po 'dis spyir chos gzhi Inga bcu tsam btsug pa las | sher 
phyin 'chad pa'i chos bzhi sum cu so Inga yod 'As soon as [Dharmapala] was reign- 
ing he invited teachers of the Prajnaparamita. He had particular faith in Acarya 
Haribhadra. This king set up about fifty religious foundations (dharmadhikarah) 
and thirty-five of them were for the exegesis of the Prajnaparamita'; HBI p. 274. 
For evidence that chos gzhi renders Sanskrit dharmadhikarah and that the latter 
means 'a religious foundation' rather than 'a centre for the Doctrine', as it is trans- 
lated in HBI p. 274 see here p. 104. 

169 Obermiller 1986, p. 156-157. For the proximity to Nalanda of the monastery 
of Uddandapura, which in Tibetan sources is known as Otantapuri, see Rgya gar 
chos 'byung, p. 156, 1. 19: o ta nta pu ri dang nye ba na na le ndra zhes bya ba'i 
gtsug lag khang zhig bzhengs 'He built the Nalanda monastery near Otantapuri'; 
HBI, p. 258. I use Uddandapura because this is what we find in a pedestal inscrip- 
tion found at Bihar Sharif in the Patna District (Choudhary 1958, p. 65; Hunt- 
ington 1984, p. 213, no. 19): deyadharmmo yam srinarayanapaladevarajye samvat 
54 sri-uddandapuravastavyaranaka-uccaputratharukasya 'This is the pious gift of 
Tharuka, son of Ucha, resident at the Great Monastery of Uddandapura, in year 
54 of the reign of Narayanapaladeva'. Bihar Sharif is indeed near Nalanda. The 
form Uddandapura also occurs in an inscription of the reign of Surapala recording 
the installation of a Buddha image in the monastery there by a monk Purnadasa 
(Choudhary 1958, p. 54). As for the Nalandamahavihara, it long predates the 
Palas. Faxian (d. before 423) describes the major Buddhist edifices in this area 
but is silent about Nalanda, which implies that if it existed it was certainly not 
an institution likely to have been home to the great names of the early Mahayana. 
The Da Tang Da Ciensi sanzang fashi zhuan, the biography of Xuanzang (ordained 
between 609 and 617; left for India in 627 or 629; studied at Nalanda; d. 664) 
written by his disciple Huili and later continued and edited by Yancong in 688, con- 
tains an account of the history of Nalanda (Beal 1914, pp. 110-113), from which 


The Saiva Age 

probability that this report is accurate is increased by the fact that he, unlike 
Taranatha, knew that Dharmapala came before not after Devapala. Taranatha 
assigns it to Devapala, probably in consequence of the aforesaid confusion, 
though he also reports a tradition that it was founded by Dharmapala's father 
Gopala, the first of the Palas. 170 

Taranatha reports that Dharmapala adopted two persons as his preceptors: 
Haribhadra and his pupil Buddhajnana. While the former was a master of the 
Prajnaparamita, the latter was a renowned authority on the Tantric system 
taught in the Guhyasamaja. 111 We are told that he performed the rituals for the 
consecration of the Vikramasila monastery and was appointed as its Vajracarya. 
We also learn that, having seen omens of the future ruin of the dynasty un- 
der Dharmapala's grandson, he persuaded the king to institute a regular fire- 
sacrifice (homah) to be performed under his guidance by the Tantric officiants 
of this monastery with the purpose of ensuring that the dynasty would be long- 
lived and consequently that Buddhism would be widely disseminated. It was 
performed, we are told, for many years at huge expense. 172 Further evidence of 

it appears that it began as a small Sangharama donated by the fourth Gupta king, 
Kumaragupta Sakraditya, who reigned from 415 to 455. It then grew through the 
addition of further Viharas until by Xuanzang's time it had become the foremost 
Buddhist structure in India, famed throughout Buddhist Asia as a centre of learn- 
ing. See the analysis of the history of the Nalandamahavihara on the basis of the 
Chinese sources in KUWAYAMA 1988, pp. 7—11. For a plan of Nalanda with its row 
of nine identical monasteries and several temples see Michell 1990, p. 246. 

170 Rgya gar chos 'byung, p. 158, 11. 7-8: rgyal po go pa la 'di 'am de wapa la'i mtshams 
su dpal o ta ntapuri'i gtsug lag khang bzhengs 'The Otantapuri monastery was built 
in the period of this king Gopala or that of Devapala'; HBI, p. 262. 

171 Rgya gar chos 'byung, p. 165, 11. 10-12: seng bzang yes shes zhabs bla mar bsten | 
shes byin dang | dpal gsang ba 'dus pas phyogs thams cad gang bar mdzad | gsang 
ba 'dus pa dang 'He served Haribhadra and [Buddhah'nanapada as his preceptors, 
and filled all the directions with the Prajnaparamita and the Guhyasamaja'; HBI, 
p. 274. See also Rgya gar chos 'byung, p. 195, 11. 12-14: bi kra ma shi lar sngags 
kyi rdo rje slob dpon chen po sangs rgyas ye shes zhabs dang | der rjes mar me 
mdzad bzang pos bstan pa bskyangs At Vikramasila [firstl the Mantra- Vajracarya 
Mahapandita Buddhajnanapada and then Diparikarabhadra protected the teaching 
[of the Buddha]'; HBI, p. 325. This figure, known variously as Jnanapada (Ye shes 
zhabs), Buddhajnana (Sangs rgyas ye shes), and Buddhasrljnana (Sangs rgyas dpal 
ye shes), is a crucial figure in the history of the Mantranaya, being the source of the 
"Jnanapada" school of Guhyasamaja exegesis and practice that was introduced into 
Tibet by Rin chen bzang po. See Blue Annals, pp. 367-374 for an account of his life 
and works, and their transmission to and in Tibet. Notable among his writings are 
the Samantabhadrasadhana (Toh. 1856) and his commentary on the Guhyasamaja 
(Toh. 1852). 

172 Rgya gar chos 'byung, p. 168, 11. 6-12: rgyal po dha rma pa la la | khyod kyi tsha 
bo'i dus nas rgyal srid jig pa'i mtshan ma yod pas | sbyin sreg gi cho ga chen po 
zhig byas na yun ring du srid zin cing | chos kyang dar bar 'gyur gsungs pas | des 
kyang dngul to la 'bum phrag dgu dang nyis stong gi yo byad phul bas | slob dpon 


Genesis and Development of Tantrism 

Dharmapala's commitment to Buddhism is found in the Nesarika grant of the 
Rastrakuta king Govinda III issued in 805, since that reveals that the ensign 
depicted on his war banner was the Buddhist goddess Tara. 173 

As for Gopala (r. c. 750-775), the father of Dharmapala, whom all our sources 
make the first of the Palas, there is no evidence in the inscriptions that he 
too was a Buddhist, unless it be his having been referred to in inscriptions of 
Narayanapala (r. c. 860-917) and Vigrahapala III (r. c. 1043-1070) as a second 
Buddha. 174 However, the Rajavyakarana claims him for the faith, saying that af- 
ter a dissolute youth he converted to Buddhism and constructed various monas- 
teries, Caityas, and temples. 175 Taranatha likewise claims that he served the 
cause of Buddhism by founding many monasteries, both in Bengal, which he 
ruled in the first part of his career, and Magadha, when he had added that great 
province to his kingdom through conquest. 176 He also recounts a legend accord- 

gtso bor gyur pa'i rdo rje 'dzin pa mams kyis lo mang por sbyin sreg mdzad 'He 
told King Dharmapala: "There are signs that from the time of your grandson on- 
wards the kingdom will be endangered. If you perform a great ritual of fire-sacrifice 
you will ensure that the reign [of your line] will endure for many years and also 
that the Dharma will be disseminated". And so [the kingl had the fire-sacrifice 
done for many years by Vajradharas led by the Acarya [Buddhajnanapada], offer- 
ing substances worth 902,000 tolas of silver'; HBI, p. 278. The ritual was evidently 
a santihomah, a sacrifice for the averting of disaster. Such rituals are generic but 
they are made to serve the specific purposes of the patron by writing these into the 
formula of intention (samkalpah) that must be recited at the opening of any such 
ritual; see Sanderson 2005a, p. 357-358 and fn. 22 in a discussion of the Tantric 
Saiva ritual commissioned by the Khmer ruler Jayavarman II (r. 802-c. 835) "in 
order that this land of Kambuja [Kambujadesa] should not continue to be a depen- 
dency of Java and so that only one king should be univeral ruler [in this region]" (K. 
235, Khmer, C 11. 71-75: vrah pada paramesvara ahjen thve vidhi leha leh kampi 
kamvujadesa neh ayatta tajava ley leh ac ti kamrateh phdai karom mvay guh taja 

173 EI 34:19, 11.35-38, at the end of an enumeration of the ensigns (\rdja\cihnani) 
siezed by Govinda III from his enemies, beginning with those of the Pandya 
and Pallava kings: pandyadesadhipan matsyam vrsabham pallavesvarat | 
. . .tdrdbhagavaHlm (em. :ti Ep.) khyatam dharmad bahgalabhumipat || ittham 
etany athanyani cihnany adaya bhubhujam \ garudahkam jagattuhgo vyadhatta 
sakalam jagat 'Thus by siezing these and other royal ensigns — the fish from the 
king of Pandyadesa, the bull from the Pallava king . . . and the famous Tara from 
Dharma[pala], the king of Bengal — [Govinda III] Jagattuhga placed the whole earth 
under [the sway of] his Garuda'. 

174 The Bhagalpur plate of Narayanapala (Hultzsch 1886), 11. 4-5 and the Bangaon 
plate of Vigrahapala III (CHOUDHARY 1958, p. 83), 11. . 3-4: sa srimdn lokanatho 
jayati dasabalo 'nyas ca gopaladevah. 

175 Manjusriyamulakalpa 53.628-631. 

176 Rgya gar chos 'byung, p. 156, 11. 18-21: sku che'i stod la bham ga la la dbang bs- 
gyur | smad la ma ga dha yang dbang du bsnungs te \ o ta nta pu ri dang nye ba 
nd le ndra zhes bya ba'i gtsug lag khang zhig bzhengs | yul chen po de gnyis su dge 
'dun gyi sde mang du btsugs te bstan pa la mchod pa rgya chen po mdzad do 'In the 


The Saiva Age 

ing to which Gopala, when not yet king, found a jewel and used it as the fee 
for Tantric consecration from an Acarya. He then successfully propitiated the 
Buddhist goddess Cunda following his instructions, 177 went to the monastery 
of Khasarpana Avalokitesvara, 178 and successfully prayed to him for kingship, 
which the deity promised he would obtain if he moved east. 

In his account of Buddhism under the successors of Gopala, Dharmapala, 
and Devapala, Taranatha gives us one more report of royal monastery building. 
But unfortunately his sources seem to have been so misinformed in their pre- 
sentation of the order and identity of these subsequent kings that it is no easy 
task to discern the reign to which this building activity should be assigned. He 
tells us that Mahapala, whom he claims to have been the son and successor of 
Mahipala, built the Uruvasa monastery, described as a branch of the monastery 
at Uddandapura, and founded Buddhist establishments at the monasteries of 
Nalanda, Somapura, and Trikatuka. 179 Taranatha has his Mahipala rule for 

first part of his life he governed Vangala. In the subsequent part he subjected Ma- 
gadha. Near Uddandapura he built a monastery called Nalendra. By establishing 
many divisions of the Sangha [in monasteries] in these two large regions he greatly 
honoured the religion [of the Buddha]; HBI, p. 258. 

177 Rgya gar chos 'byung, p. 155, 1. 14-156, 1. 18; HBI, pp. 257-258. Cunda, though she 
appears not have been a major constituent of learned Tantric Buddhism, seems to 
have been popular in the region. Two bronze statues of this goddess have been found 
in Pala territory, one from Kurkihar cast in the reign of Mahipala I, and the other 
from Nalanda, assigned by Huntington on stylistic grounds to the ninth century 
(Huntington 1984, pp. 60-61, 226-227, and 144; figs. 61 and 169; wrongly giving 
the name as Cunda); and there was a temple of Cunda in Pattikera (Mainamati) 
near Comilla, which is illustrated in a manuscript of the Astasahasrika Prajna- 
pdramitd (ULC MS Add. 1643, copied in 1015), as one of eighty-five illustrations of 
Buddhist sacred sites, most in eastern India, with the legend pattikere cunddvara- 
bhavane cunda (MlTRA 1971, p. 244). There are images of Cunda from Ratnagiri, 
Udayagiri, and Achutrajpur in Orissa, Ellora in Maharashtra, Gujarat, and Nepal; 
see Shaw 2006, pp. 265-274; IAR 2001-02, Plate 114 (Udayagiri). 

178 In HBI (p. 257) it appears as "the temple of drya * Khasarpana". But the Tibetan 
states that it was a monastery: 'phags kha sa rpa na'i gtsug lag khang (Rgya gar 
chos 'byung, p. 155, 11. 20-21). A Khasarpana located in Radha is mentioned in the 
Zhib mo rdo rje of Dmar ston Chos kyi rgyal po (c. 1198-1259) as very famous in the 
time of 'Brog mi, who died c. 1064 (Blue Annals, p. 72); see Zhib mo rdo rje, p. 86, 
§4: rgyar gar shar phyogs ra da na 'phags pa spy an ras gzigs dbang phyug 'khar sa 
pa ni bzhugs pa de grags pa che pas .... Perhaps this was the site of the monastery 
referred to here. 

179 Rgya gar chos 'byung, p. 175, 1. 2-7: o ta nta pu ri'i gtsug lag khang du nyan thos 
kyi dge 'dun mams gtso bar mchod cing \ dge slong Inga brgya dang chos ston pa 
Inga bcu la 'tsho bo sbyar | de yi Ian yag tu u ru ba sa zhes bya ba'i gtsug lag khang 
bzhengs | deryang nyan thos pa se ndhapa Inga brgya re la 'tsho ba sbyor | bi kra ma 
shi lar sngargyi srol de kagzung ste | mchod 'os kyi mthil du mdzad | dpal na la nddr 
yang chos gzhi 'ga' re btsugs | so ma pu ri dang | na le ndra dang | tsha ba gsum kyi 
gtsug lag khang la sogs par yang chos gzhi mang po btsugs '[Mahapala] honoured 
principally the community of Sravakas in the Uddandapuravihara and [there] pro- 


Genesis and Development of Tantrism 

fifty-two years and says that he died at about the same time as the Tibetan king 
Khri ral, 180 that is to say, Khri gtsug lde brtsan also called Ral pa can, who 
ruled from about 815 to 836; and his son Mahapala is assigned a reign of 41 
years, 181 that is to say, up to about 900. Now, there are two Mahipalas known to 
us from the epigraphical record, both of whom were much later, the first ruling 
c. 977-1027 and the second c. 1070-1071; but there is no Mahapala. The similar- 
ity with the name of his father raises the suspicion that one king Mahipala, no 
doubt Mahipala I, the length of his reign agreeing closely with that attributed to 
Mahipala by Taranatha, has become Mahipala and Mahapala, and that the re- 
sulting two reigns, amounting implausibly to ninety-three years, served to bridge 
a gulf of ignorance of the period between the great founders of the Pala empire 
and Mahipala I, who restored the fortunes of the Palas after a period during 
which, following Devapala, they had lapsed into insignificance, losing control of 
Bengal and retreating into a core territory in Bihar around modern Patna. 182 It 
is probable, then, that Taranatha's attribution to Mahapala of the expansion of 
Uddandapura and the founding of Buddhist establishments at Nalanda, Soma- 
pura, and Trikatuka is a distortion of a record of the pious works of Mahipala I. 
The supposition is somewhat strengthened by the fact that Taranatha says that 
the Kdlacakratantra was introduced during the latter half of Mahipala's life and 
that it spread during the reign of Mahapala. 183 For it was during the reign of 
Mahipala I that this new Tantric system emerged. 184 

vided for five hundred monks and fifty teachers of the Dharma. As a branch of this 
he built a monastery called Uruvasa. In this too he provided for five hundred Saind- 
hava Sravakas. He accepted that the pre-existing system at Vikramasila should 
remain unchanged; but he made [Uruvasa] the object of his greatest veneration. He 
also established several religious foundations at Nalanda, and many others also in 
Somapura, Nalendra, and the Trikatukavihara'; HBI, p. 289. 

180 Rgya gar chos 'byung, p. 172, 11. 1-3: de nas rgyal po ba na pa la'i sras ma hi pa la 
zhes pa byung | rgyal srid lo Inga bcu nga gnyis mdzad | rags rtsis su byas na rgyal 
po 'di 'das tsam na | bod na btsan po khri ral yang sku 'das pa tsam gyi dus yin no 
'Next, the son of Vanapala, called Mahipala, ruled for fifty-two years. By a rough 
calculation this king died at the same time as King Khri ral in Tibet'; HBI, p. 284. 

181 Rgya gar chos 'byung, p. 175, 1. 1: de 'i sras ni rgyal po ma ha pa la ste | 'dis rgyal 
srid lo bzhi bcu zhe gcig mdzad 'His son was King Mahapala. He ruled for forty-one 
years'; HBI, p. 289. 

182 See Smith 1962, pp. 412-418; and Kulke in Kulke and Rothermund 1992, 
p. 118. 

183 Rgya gar chos 'byung, p. 175, 11. 7—9: rgyal po ma hi pa la'i sku tshe'i smad tsam na | 
pi to a tsa ryas dus kyi 'khor lo'i rgyud spyan drangs te | rgyal po 'di'i dus su dar bar 
mdzad 'The Acarya Pito introduced the Kdlacakratantra in the second half of the 
life of King Mahipala and disseminated it during the time of this king [Mahapala]'; 
HBI, pp. 289-290. This Pito is no doubt the person elsewhere called Pindo (Bsod 
nyoms); see Blue Annals, p. 756-757, 789; OROFINO 1994, p. 23. 

184 Newman 1987 and 1998; Orofino 1994, p. 23. 


The Saiva Age 

After Mahipala the monastic universities already established continued to 
flourish, but Pala fortunes once again went into decline, and it is therefore not 
surprising that Taranatha has no major royal benefactions to report during this 
period. However, during the long reign of Ramapala (r. c. 1072-1126), the last 
major ruler of this dynasty, the kingdom recovered, and we might expect this to 
be reflected in a renewal of material patronage. It is tempting therefore to accept 
the claim made by Hara Prasad Shastri in 1910 185 and repeated by many since 
that time 186 that the Jagaddalamahavihara, 187 the one great monastery in the 
Pala domains whose founder has not yet been identified, was the creation of this 
monarch. But there is no evidence that supports this claim 188 

Nor is there any that refutes it. In the introduction to the edition of the 
Subhasitaratnakosa published by KOSAMBI and GOKHALE the former has 
asserted on the strength of evidence provided by the latter that Ramapala's 
coronation took place in this monastery, 189 in which case, of course, it could not 
have been founded by him during his reign. But that too cannot be accepted. 
The evidence cited is Gokhale's rendering of the colophonic verse at the end of 
the *Bhagavatyamnyayanusarini vyakhya, a commentary on the Astasahasrika 
Prajnaparamita which survives in Tibetan translation (Toh. 3811): 190 "This 
vyakhya was composed by Raja-jagaddala-nivasi [which thus becomes the 
writer's name] at the Jagaddala vihara, which was the place of Ramapala's 
coronation". 191 But this rendering is wildly inaccurate. The meaning of the 
Tibetan is: "I, a resident of the venerable Rajajagaddala [monastery], have 
composed this commentary, a string of pearls (muktavali) [to be an adornment] 
of the land protected by King Ramapala". 192 This does at least convey the 

185 Ramacarita of Sandhyakaranandin, introduction, p. 9. 

186 E.g. Mookerji 1951, p. 595; Rahul Sankrityayana cited by Kosambi in 
Kosambi and Gokhale 1957, p.xxxviii; Krishnamacharya, p. 1 of his San- 
skrit introduction to Tarkabhdsd (1942); MlTRA 1971, p. 16; cf. HUNTINGTON 1984, 
p. 196. 

187 It is referred to as a Mahavihara in the colophonic verse of Munisribhadra's Panca- 
kramatippanl (munisribhadrena ciraj jagaddalamahdviharasadbhiksuna) and in 
3.7 of the Ramacarita of Sandhyakaranandin (jagaddalamahaviharacitaragam). 

188 Kajiyama 1998, p. 7. 

189 Subhasitaratnakosa, p. xxxvii, fn. 8. 

190 bCom Idan 'das ma'i man ngag gi rjes su 'brung ba zhes bya ba'i mam par bshad 
pa, f. 320r2: mi yi dbang po ra ma pa las sa skyong mdzad pa'i <gnas kyi [Cone, 
Peking]> mu tig phreng ba ni | dpal Idan rgyal po dza ga ta la gnas par byed pa 
bdag gis mam bshad 'di byas so. 

191 Subhasitaratnakosa, p. xxxvii, fn. 8. 

192 Gokhale seems to have found his "coronation" in the dbang of mi yi dbang po 
ra ma pa las. The word is used in Tantric texts as a short form for dbang bskur 
'consecration' (abhisekah), as at rGyud spyi, p. 270, 1. 1. But in order to reach his 
understanding of the phrase in which it occurs he has had to forget the mi yi that 


Genesis and Development of Tantrism 

valuable information that the monastery was a royal foundation, since the 
Tibetan of its name dpal Idan rgyal po dza ga ta la, is evidently a translation 
of srlmadrajajagaddala- , a form of the name confirmed by its occurrence in 
Sanskrit at the end of Moksakaragupta's Tarkabhdsd, in which he informs us 
that he too was a resident of this monastery (snmadrajajagaddalavihariya-). 193 
But we remain ignorant of the king who founded it. We know that it existed in 
the time of Ramapala, and it is not impossible that it was indeed the work of this 
last great king of the dynasty; but no evidence of which I am aware precludes its 
having been created by a predecessor. 

Some idea of the scale of the Great Monasteries in the Pala domains is pro- 
vided by Taranatha. He informs us that in the reign of Ramapala, even after 
the decline from the time of the early Palas, there were one hundred and sixty 
monks holding posts as Panditas at Vikramasila, and that there were about 
a thousand monks permanently in residence, both there and at Uddandapura, 
with many more assembling on the occasion of festivals. 194 We also learn that 
when Vikramasila was founded its design incorporated one hundred and eight 
shrines: a central temple housing a life-size statue of the Great Awakening 
(Mahabodhi) 195 surrounded by fifty-three small temples dedicated to the inner 

precedes — mi yi dbang po 'king', lit. 'lord of men', rendering Sanskrit nrpatih, 
narendrah, or a synonym — , the fact that ra ma pa las after it is instrumental 
not genitive, and the fact that the emphatic and separative particle ni that ends 
the larger phrase of which this is part and marks it out as the subject militates 
against its being taken as qualifying the monastery. The expression mu tig phreng 
ba describing the commentary figuratively as a string of pearls is probably also in- 
tended to convey its title by paronomasia, i.e. Muktavali, a title found elsewhere 
in this literature, for example as the title of Ratnakarasanti's commentary on the 
Hevajratantra. The author remains anonymous. 

193 Tarkabhasa, p. 39. Kajiyama (1998, pp. 6-11) shows that Moksakaragupta was 
active at some time after c. 1050 and before c. 1292. 

194 Rgya gar chos 'byung, p. 189, 11. 13-19: bi lira ma shi lar pa ndi ta brgya drug cu 
tsam re dang | gtan du du bzhugs pa'i dge slong stong re yod cing | mchod pa la sogs 
pa'i dus su rab byung Inga stong re 'du | rdo rje gdan du rgyal pos tsho ba sbyar ba'i 
theg chenpa bzhi bcu re dang | nyan thos kyi dge slong nyis brgya re rtag tu bzhugs 
shing | dus dus su nyan thos kyi dge slong khri phrag re tshog pa byung | o ta nta 
pu rir yang rtag tu dge slong stong phrag re bzhugs | theg pa chen chung gi ste gnyis 
char yod cing | dus dus su rab tu byung ba mams 'dus pa stong phrag bcu gnyis re 
'byung bar grags 'There were at least 160 Panditas in Vikramasila and 1000 monks 
who were permanent residents. As many as 5000 renunciate monks gathered there 
on the occasion of festivals and the like. At Vajrasana (Bodhgaya) 40 adherents 
of the Mahayana and 200 Sravaka monks resided permanently, maintained by the 
king. From time to time as many as 10,000 Sravaka monks congregated there. In 
Uddandapura there were 1000 permanently resident monks, comprising adherents 
both of the Mahayana and of the Hmayana. From time to time 12,000 renunciate 
monks gathered there'; HBI, p. 313. 

195 I take this to be an image of Sakyamuni attaining enlightenment seated beneath 


The Saiva Age 

deities of the Mantranaya (gsang sngags nang gi lha khang chung ngu) and 
fifty-four "common" temples {lha khang dkyus ma), that is to say, temples en- 
shrining exoteric, non-Tantric images. The king, we are told, provided generous 
allowances for the food and clothing of one hundred and eight Panditas, three 
Vajracarya specialists to perform Bali offerings, rituals of image-installation, and 
fire-sacrifices respectively, and three officials. The first is the 'Guardian of Duties' 
(bya ba bsrung pa), perhaps an official appointed to ensure monks' adherence to 
the various roles assigned to them in the running of the monastery. The second 
is termed mysteriously 'Guardian of Doves' (phug ron bsrung pa), and the third 
is the 'Supervisor of the Monastery's Subjects' {lha 'bangs kyi gnyer byed pa), 
these being, perhaps, both the serfs or tenants that worked the monastery's es- 
tates and the servants within the monastery itself. 196 Archaeological excavations 
have revealed that the cell-lined square court of Vikramasila 197 measured 1073 
feet on each side, that the entire site was spread over an area of more than one 
hundred acres, 198 and that Dharmapala's monastery at Somapura (Paharpur) 
was of similar design and plan and of only slightly smaller size, 199 as was the 
monastery founded by Bhavadeva of Samatata at Pattikera (Mainamati). 200 We 
also have some information concerning the scale of the monastery at Nalanda 
during the early seventh century when the Chinese scholar Xuanzang was there. 
According to the account written by his pupil Huili there were as many 10,000 

the Bodhi tree, as in the case of the approximately contemporary principal image in 
the central shrine of Monastery 1 at Ratnagiri, though that is somewhat larger than 
life-sized, the figure seated in the lotus posture being over two metres in height. See 
Harle 1994, p. 163; Huntington 1985, fig. 19.44. We see another example in the 
central shrine at Udayagiri (IAR 1997-98, Plate 101; 1998-99, Plate 48). 

196 Rgya gar chos 'byung, p. 165, 1. 17-p. 166,5; HBI, p. 275. The three specialists are 
a gtor ma'i slob dpon, a rab gnas slob dpon, and a sbyin sreg slob dpon, i.e. a 
balyacaryah, a pratisthacaryah, and a homacaryah. 

197 On the reasons for identifying the monastery at Antichak with the Vikramasila- 
mahavihara see p. 88. 

198 Mitra in EITA, vol. 2, pt. 2, p. 403; IAR 1972-1973, pp. 4-5 (the western outer wall 
shows a length of 330 metres; p. 5 gives a plan of the excavated structures); IAR 
1973-4, pp. 8-9 (northern wall measures 330 metres). 

199 Dikshit 1938, pp. 18-36. Plate I (general plan). He reports (p. 18) that the outer 
quadrangle measures 822 feet externally on each side (according to Mitra in EITA, 
vol. 2, pt. 2, p. 403, it measures 922 by 919 feet) and (p. 34) that the original 
monastery was designed to accommodate some 600 to 800 monks and that in the 
eleventh century the number of residents can have been no more than 400. The 
massive central cruciform shrine-complex measures 386 by 352 feet. 

200 This monastery is probably that known as the Salban Vihara, consisting like the 
monasteries of Vikramasila and Somapura of a massive cruciform shrine within a 
square enclosure which though considerably smaller than that of those monasteries 
was nonetheless of great size, each side being 550 feet in length; see Mitra inEITA, 
vol. 2, pt. 2, pp. 402-403. 


Genesis and Development of Tantrism 

monks there, all Mahayanists, either as permanent residents or visitors, and 
over a 1000 learned scholars. 201 

These royal monasteries are likely to have accumulated great wealth. The 
tax-exempt agricultural lands granted to them at the time of their foundation 
would have provided them with a substantial initial endowment: Huili reports 
that Nalanda's was the revenue of about 100 villages; 202 and the wealth from this 
source would no doubt have been augmented by subsequent land-grants 203 and 
would certainly have been augmented by other votive donations, bequests from 
the estates of deceased laymen, 204 and the profits of such non-religious activities 
as banking and the provision of irrigation and other agricultural facilities. 205 

No doubt they would also have benefitted from the riches accumulated by 
individual monks in the form of the rewards (daksina) that they earned by giving 
initiations, imparting instruction, installing images, consecrating monasteries 
and temples, reciting sacred texts, and performing rites for protection, funeral 
ceremonies, and the like. 206 Tibetan sources record the very large amounts of 
gold which Indian and Tibetans required for such services. 'Brog mi agreed to 
give the Indian Gayadhara 100 gold srang, some 3,750 grams, each year for five 
years in return for the transmission of the esoteric Lam 'bras teachings; 207 Zur 
po che sha kya 'byung nas offered 'Brog mi 100; 208 Rva lo tsa ba gave 100 srang 
to the Nepalese Guru Bha ro phyag rdum for the Yamdri cycle instructions; Se 

201 Beal 1914, p. 112 

202 Beal 1914, p. 112 


We have a record (EI 17:17: the Nalanda copper-plate of Devapala) of one such sub- 
sequent land-grant in the case of the monastery at Nalanda. This records that in 
the 35th year of Devapala, c. 847, five villages were assigned for the support of the 
Buddha, the Dharma, and the Sangha of a new monastery (vihdrah) constructed at 
this site by Maharaja Balaputradeva, the Sailendra king of Suvarnadvipa (Suma- 
tra). That the regnal year is the 35th is the view of Sircar (1983, p. 79, note 38). 
Hirananda Shastri read the numerals as 39 (EI 17:17, 1. 42). 

204 The Mulasarvdstivddavinaya speaks of the validity of written wills in which 
wealthy laymen transfer their entire estate to the the Sangha; see Gilgit 
Manuscripts vol. 3, pt. 2, p. 140, 1. 14-15, 1. 1; and SCHOPEN 2004, p. 6. It also 
sets out rules obliging monks to accept permanent endowments of cash (aksayanivt) 
(Schopen, loc. cit.). 

205 On the profit-making activities of Buddhist monasteries in the fifth and sixth cen- 
turies in India and in China under the Northern Wei (386-534) see Liu 1994, pp. 
120-158. As for banking, the Mulasarvdstivddavinaya requires the funds of per- 
manent endowments (aksayanivt) for the Buddha, the Dharma, and the Sangha to 
be lent out on interest (vrddhih) (SCHOPEN 2004, pp. 6-7, 47-49, 53). On monas- 
tic landlordism and the profitable management of irrigation works, in which local 
farmers were given access to such facilities in return for a share of their crops as a 
donation to the Sangha see Shaw and Sutcliffe 2003 and Gunawardana 1979. 

206 p or ^g daksina for the Tantric funeral ceremony see here p. 102. 

207 Zhib mo rdo rje, p. 90, Blue Annals, p. 207 

208 Zhib mo rdo rje, p. 92 


The Saiva Age 

tsha bsod nams rgyal mtshan gave 50 srang to the Nepalese KayasrI for the 
precepts of the Nam mkha' skor gsum; Mar pa performed a rite to protect the 
sons of some wealthy men and charged 10 gold srang for each son; 209 and the 
hagiographies of early Tibetans who travelled to India to acquire initiation and 
instruction abound in reports of the need to amass large quantities of gold for 
this purpose. 210 

It would be rash to assume that the fortunes that were garnered in this 
way by Indian Acaryas were added directly to the resources of their monasteries. 
A passage in the Mahavairocanabhisambodhitantra, a text produced in the sev- 
enth century, at the beginning of the history of the Mantranaya as a fully-fledged 
path within the Mahayana, 211 suggests that this was the case: 212 

After the \santikd\homah the Mantrin should request from the disciples a fee 
(daksina) of gold, silver, jewels, a stallion, an elephant, a mare, a cow, a bull, a 
buffalo, cloth, and whatever else is fitting. At that time the disciples should give 
the daksina to the Guru, respectfully, with faith, generating joy in their minds. 
Or at any rate they should make the Guru entirely satisfied. After [the Mantrin, 
that is to say, the Guru] has done this he should do a rite of self-protection and 
then exhort the excellent disciples as follows: All the Buddhas teach that this is a 
field for [the sowing of] merit for the benefit of all living beings. Therefore give to 
the Sahgha, [for it is] vast in its pure virtues. 

But it is striking that references to the Sarigha are not found in this context in 
later texts, which only specify the goods that should be given.These are much the 
same as in the Mahavairocanabhisambodhi, though Diparikarabhadra, setting 
out the procedure for initiation with the Mandala of the Guhyasamaja, adds land 

209 Blue Annals, pp. 377, 395, and 400. 

210 See, for example, pp. 399-401 of the account of the life of Mar pa in the Blue Annals. 

211 The earliest certain evidence of the text is its Chinese translation by 
Subhakarasimha and Yijing registered in A.D. 725 (TaishS 848). But Hodge 
(2003, pp. 14-15) points out that Yijing's Xiyuqiufaguosengzhuan ('Record of Em- 
inent Monks who Sought the Dharma in the West') reports that the monk Wuxing, 
his contemporary in India, had died as he was setting out to return to China in 
674, that texts he had collected were forwarded to China, and that three important 
Tantras are listed among these works: the Subahupariprccha, the Susiddhikara, 
and the Mahavairocanabhisambodhi. 

212 rNampar snang mdzad chenpo mngon par byang chubpa'i rgyud, f. 173r4— 7: sbyin 
sreg rjes la sngags pa yis | slob ma mams la yon bslang ba | gser dang dngul dang 
rin chen dang | rta dang de bzhin glang po dang | rta mo ba lang ma he gos | gzhan 
yang dngos po ci yang rung | de tshe slob mas gus par ni | dad pa rab tu Idan pa yis 
| sems la dga' ba bskyed nas su | bla ma ni yon bdul lo | yang na ci nas bla ma de | 
rab tu mgu bar 'gyur bar bya \ de Itar byas nas bdag bsrung ste \ slob ma de pos bsgo 
ba ni \ 'di ni bsod nams zhing yin zhes | sems can kun gyi don gyi phyir | skyob pa 
mams ni kun gyis gsungs \ mam dag yon tan rgyas pa yi \ dge 'dun la ni kun gyis 
by in. 


Genesis and Development of Tantrism 

at the head of the list, 213 and the scripture Laghusamvaratantra goes so far as in- 
clude a rastram, which I take to mean [the revenues of] 'a district' or 'sub-district' 
of a kingdom and therefore to be envisaging the gift of a monarch. 214 Moreover, 
the Mulasarvastivadavinaya, which was the predominant code of monastic law 
in eastern India and was thence adopted in Tibet, recognizes that monks had 
private property and that there could be great differences of wealth owned by 
individuals within the Sarigha. However, it also insists that such property does 
not go to the king when a monk dies, as brahmanical law required in the case of 
those who die without offspring, but remains within the monastic community to 
which he belongs. 215 Of course, a wealthy Guru could also donate his wealth to 

213 Guhyasamajamandalavidhi, f. 16vl-2, v. 375c: bhugajadisuvarnadau 'land, 
an elephant or [other mount], gold, and other [valuables]'. The Mrtasugati- 
niyojana of Sunyasamadhivajra includes houses, land, and male and female 
slaves among the gifts that should be given to an officiant who performs 
the Tantric funerary ceremony (antyestih): yojanako 'pi svavibhavdnurupam 
vastralamkarasayanasanagrhaksetradasTdasadikam daksinam acarydya sadaram 
dadyat (f. 4r2-3). 

214 Laghusamvara f. 4rl-3 (3.11-14b): tatas tu gurave dadyat tathagatoktadaksinam 
| nirjatyam suvarnasatasahasram ratnani vividhani ca || 3.12 vastrayugmasatam 
caiva gaja vajT rastram eva ca | karnabharana katakam ca kanthikdhgulikais ca 
samuttamam || 3.13 yajhopavita sauvarnam svabhdrydm duhitam api | ddsa dasi 
bhagnim vdpi pranipatya nivedayet 'Then he should give to the Guru the daksind 
prescribed by the Tathagata. After prostrating himself he should give 100,000 
[Palas] of the most precious gold, jewels of various kinds, 200 lengths of cloth, an 
elephant, a horse, and a rastram, earrings, bracelets, necklaces, rings, and a crown, 
a golden caste-thread, his wife, his daughter, a male slave, a female slave, or his 
sister'. The use of the term rastram for 'a district' or 'sub-district' is seen in in- 
scriptions; see Sircar 1966, pp. 277-278. My translation of the passage follows the 
text and interpretation of the commentator Bhavabhatta. The reading nirjatyam, 
which he interprets as 'most precious', is suspect. The MS (Laghusamvara, f. 
4r2) reads the much more satisfactory niryatya 'having given', as does the com- 
mentator Kambalapada (Sddhananidhi, f. Ilv4); and this is also the reading seen 
in f. 54v3-5 of the Samvarodayd ndma mandalopayika of Bhuvacarya of Ratna- 
giri in Orissa (see here p. 91), in the Nepalese codex unicus of 1056. See also 
Catuspithatantra f. 60vl-2 (4.1.46-47), which includes a house, land with rights 
to mine, and grain: tato gurudaksinam dadya sisya bhdvena nityasah \ dtmapatnim 
saputram vd bdndhavaih saha cetikaih | hasti asva gavadinam grha ksetras ca go- 
travan || sauvarna rajata tamram vastradi vrihidhanyakaih . The Vimalaprabha 
on Kdlacakratantra, Abhisekapatala v. 198 explains that verse as meaning that the 
initiate should promise always to give to his Guru one sixth of all his inherited and 
self-acquired wealth in the form of gold, jewels, grains and the like, and a sixth of 
all his livestock. It adds that he is required to give his wife to the Guru five times 
each month (vol. 2, p. 144, 11. 17-22). 

215 The inheritance of the property of deceased monks is treated in the Mulasarvasti- 
vadavinaya in the Civaravastu (Gilgit Manuscripts vol. 3, pt. 2, pp. 113—148). Par- 
ticularly relevant in this context is its discussion of the case of the monk Upananda, 
who died leaving 300,000 in gold (pp. 117-121). King Prasenajit is persuaded that 
the estate does not belong to the crown and the Buddha rules that it should be 


The Saiva Age 

the monasteries during his lifetime by creating religious endowments. We have 
a striking example of this in the eleventh century. Rva Lo tsa ba, who had be- 
come extremely wealthy by charging for instruction in the Tantras — he is said 
to have established fixed rates for a wide range of texts — , sent 100 srangs of 
gold to Vikramasila to fund the recitation in perpetuity of a copy of the Panca- 
vimsatisahasrika Prajnaparamita written in gold, two golden copies of the Asta- 
sahasrika Prajnaparamita, and 100 srangs of gold to fund the recitation in per- 
petuity of eighty -four copies of the Prajnaparamitasamcayagatha by eighty-four 
Panditas of the monastery. 216 

How closely the Pala emperors and their bureaucracy were involved in the 
supervision of their Buddhist foundations cannot be determined from the avail- 
able evidence. But it is almost certain that a Superintendent would have been 
appointed by the ruler to oversee their administration and that he would have 
required a substantial staff to enable him to do so. The Ratnavali, a Mahayanist 
work of uncertain authorship written before the sixth century 217 advises the un- 
known king to whom it is addressed on the proper administration of his realm 

distributed among the monks of his monastery: bhajayata yuyam bhiksava upa- 
nandasya bhiksor mrtapariskaram (p. 119, 11. 13—14). The main concern here is to 
ensure that the wealth of monks stays within the community, free of the state's 
interferecee. For analysis of the treatment of these and related matters in the 
Mulasarvastivada-vinaya see SCHOPEN 2004, pp. 3-6. The private property of a 
deceased monk was to be divided, directly or after sale, among the members of his 
community or, where this was not appropriate, as in the case of land, servants, and 
grain-stores, taken over for the use of the whole community (Gilgit Manuscripts, 
vol. 3, pt. 2, pp. 141, 1. 4-143, 1. 1). But when the estate contained precious metals, 
worked or not, those were to be divided into three shares, one for each of the Three 
Jewels (Gilgit Manuscripts, vol. 3,pt. 2, p. 143, 11.10-12: suvarnam ca hiranyam 
ca yac canyac ca krtakrtam trayo bhagah kartavyah | eko buddhasya | eko dhar- 
masya trtiyah sahghasya). That for the Buddha should be used for repairs to the 
monastery's Buddha shrine (gandhakuti) and relic Stupas, that for the Dharma 
should fund the copying or enthroning of the Buddha's teachings, and that for the 
Sahgha should be divided among the monks (ibid., 11. 12-14). In the case of jewels 
other than pearls half should go to the Dharma and half to the Sarigha (ibid. , 11. 1- 
5). Manuscripts of Buddhist texts should be added to the monastery's library and 
manuscripts of non-Buddhist texts should be sold and the proceeds shared (ibid. , 
11. 5-7). 

216 Blue Annals, p. 377. 

217 The work is attributed to the Nagarjuna of Mulamadhyamakakarika fame. I con- 
sider this attribution to be doubtful in spite the fact that it is made by such au- 
thors of the sixth century and later as Bhavaviveka, Candrakirti, Haribhadra, Ka- 
malasila, and *Ajitamitra (Mi pham bshes gnyen), who wrote the only known com- 
mentary on the text, which has come down to us in a Tibetan translation made by 
the Bande Dpal brtsegs with the Indian Vidyakaraprabha in the early ninth cen- 
tury. The Ratnavali itself contains no evidence of its authorship and Vetter (1992) 
has cast doubt on the traditional attribution through an analysis of its metre and 
word frequency. 


Genesis and Development of Tantrism 

and begins by declaring: "Appoint for all religious foundations a Superinten- 
dent of Religion {dharmadhikrtah) who is energetic, without avarice, learned, 
and virtuous, who will not oppress them". 218 It goes on to advise him on the 
qualities he should look for in those whom he appoints as ministers (sacivah), 
military commanders (dandanayakah), and superintendents (adhikrtah), telling 
the king: "Have them submit to you complete monthly accounts of revenues and 
outgoings and, after hearing these, personally conduct all business pertaining to 
religious foundations and the rest". 219 This, of course, is not evidence of what 
was done in the Pala realm. But as I read the passage it is the qualities and 
duties of these various officials that are the subject of injunction, not their exis- 
tence; and there is certainly nothing exceptional in the office itself, since we have 
evidence that it was normal in kingdoms throughout the Indie world. 220 Ab- 

218 Ratnavali 4.22: sarvadharmddhikdresu dharmddhikrtam utthitam \ alubdham 
panditam dharmyam kuru tesdm abddhakam. The term dharmddhikdrah, which 
elsewhere is used to refer to the office of the Superintendent, is clearly used here in 
the meaning 'religious foundation', as the Tibetan translation chos kyi gzhi agrees, 
and as it occurrence earlier in the same passage (4.18) confirms: dharmddhikdrd ye 
cdnye purvardjapravartitdh | devadronyddayas te 'pi pravartyantdm yathd sthitdh 
'And you should ensure that temples and other religious foundations created by 
former kings should continue as they are'. This sense of the word is also found 
in Licchavi inscriptions; see LKA 71, 11. 12; and 81, 1. 11-12: bhavisyadbhir 
api bhupatibhih purvardjakrtadharmddhikdrapdlanddrtair bhavitavyam 'Future 
kings too must take care to maintain religious foundations created by kings of the 

219 Ratnavali 4.26: pratimdsam ca tebhyas tvam sarvam dyavyayam srnu | srutvd 
*dharmddhikdrddyam kdryam sarvam (Tib. chos gzhi sogs kyi don kun nyid) 
svayam kuru. 

220 In the Abhijndnasdkuntala of Kalidasa Dusyanta, wishing to conceal his identity 
from Sakuntala tells us that he has been appointed by the king to the office of Su- 
perintendent of Religion and accordingly has come to her hermitage in his official 
capacity to satisfy himself that they are free of hindrances to the performance of 
their rites; Act 1, after v. 22, p. 38: bhavati yah pauravena rdjnd dharmddhikdre 
niyuktah so 'ham avighnakriyopalambhdya dharmdranyam dydtah. The fifth Da- 
modarpur copper-plate inscription, of 533/4, recording a formal request for the pur- 
chase of land in the Kotivarsa district to be given to a nearby temple, speaks of it 
being presented with the full knowledge of the Office of Religion (dharmddhikdra- 
buddhya) {EI 15:7, p. 143). A banker Ralhana has the title dharmakarmddhikdrl 
'the superintendent of religious activities' in the Kharod inscription dated in 1181/2 
of Ratnadeva III, the Kalacuri of Ratnapura (EI 21:26, 1.28: sresthind ralhane- 
ndtra dharmakarmddhikdrind). The humourous play Agamadambara, composed 
by the Kashmirian philosopher Jayantabhatta and set in the Kashmir of his own 
time, during the reign of Sankaravarman (883-902), has a Saiva ascetic inform us 
that a brahmin Samkarsana has been appointed by that king to the dharmaraksd- 
dhikdrah, the 'Office of Superintendent of Religion' for the whole country (Act 3, 
Prelude, p. 132: sakalde yyeva vasumdhalde dhammalaskddhidle niutte [*sakaldyd 
eva vasumdhardyd dharmaraksddhikdre niyuktah]). The term dharmadhikrtah oc- 
curs in a fifteenth-century inscription from Nilacala, the site of the famous temple 


The Saiva Age 

sence of thorough external control of the great monasteries seems all the more 
unlikely when one considers that apart from the fact that they were such large 
and wealthy establishments it was not the case that by building, equipping, and 
endowing a monastery a patron surrendered his ownership entirely The patron 
continued to be the owner of the monastery and its contents (mahaviharasvami, 
viharasvami) in some sense and the monks were obliged to employ all these for 
the purposes for which they were designated, the return for the owner being the 
constantly augmenting merit that was generated for him by their repeated use 
(paribhoganvayam punyam). Only where there was no such use, as in the case of 
a Caitya, did a donor gain merit once and for all by the simple act of surrendering 
ownership (tyaganvyayam punyam). 221 

Moreover, we know that monks who held senior teaching positions in the 
great monasteries did so by royal appointment, 222 and that rituals for state pro- 

of Kamakhya, near Gauhati in Assam, recording a grant of land by a king Madhava. 
The inscription opens with the information that the grant has the approval of this 
official: dharmadhikrtenanumatam (SlRCAR 1979, p. 16, 1. 1). Mpu Prapanca re- 
veals in his Old Javanese poem Desawarnana that there were two Superintendents 
of Religion in the Majapahit kingdom of east Java, one for the Buddhists (dharma- 
dhyaksa kasogatan), and the other for the Saivas (dharmadhyaksa kashaiwan). 
Inscriptions from that kingdom reveal that there was also a board of subordinate 
religious officials known as the Assessors of Religion (dharmopapatti or dharmadhi- 
karana); see Santiko 1995, p. 56; cf. here p. 119; for references see ZOETMULDER 
1982, under dharmadhyaksa, dharmopapatti and dharmadhikarana. 

221 On this crucial distinction between paribhoganvayam punyam and tyaganvayam 
punyam see Vasubandhu, Abhidharmakosabhasya on 4.121a (caitye tyaganvayam 
punyam 'In the case of a Caitya there is merit that accrues from surrender'): caitye 
saragasyatmartham danam ity uktam | tatrasaty upabhoktari katham punyam bha- 
vati | dvividham punyam tyaganvayam tyagad evayad utpady ate paribhoganvayam 
ca deyadharmaparibhogad yad utpadyate | caitye tyaganvayam punyam (4.121a) 
'It has been said that a gift to a Caitya made by one who is not free of attachment 
is for his own benefit. Since there is no enjoyer of the gift in such cases how can 
there be merit [generated by such a gift]? Merit is of two kinds: tyaganvayam, 
which arises only from the surrender [of ownership of what is given], and parib- 
hoganvayam, which arises from the enjoyment of a pious gift [by the recipients]'. 
One should note that the restrictive particle eva is used here only after tyagad. Va- 
subandhu does not state conversely in the case of paribhoganvayam punyam that 
this kind of merit arises only (eva) from the use of the donation. I infer that merit 
in such cases was understood to arise both from the act of surrendering possession 
and from subsequent use. This is confirmed by Candrakirti, who in his Prasanna- 
pada, commenting on paribhoganvayam in Madhyamakakarika 17.5a, speaks of 
the goods used as 'surrendered' (parity aktasy a). See Abhidharmakosabhasya on 
4.4ab addressing the conundrum of how the Buddha's doctrine of moral action as 
intention (cetana) can be reconciled with this claim of the accretion of further merit 
(punyavrddhih) whenever a recipient uses something donated whether or not the 
donor is aware of it; and Sanderson 1995c, pp. 38-40. 

222 Rgya gar chos 'byung, p. 179, 11. 13-14: rgyal pos spyan drangs te na la nda dang 
| bi kra ma la shi la'i nub sgo bar bskos shin 'The king invited [Vaglsvarakirti] 


Genesis and Development of Tantrism 

tection were performed on behalf of the monarch at Vikramasila. We have seen 
above Taranatha's report of the fire-ritual performed for the benefit of the dy- 
nasty by the Vajracaryas of that monastery; and two important texts on the rit- 
ual of initiation written by two major Tantric authorities under the early Palas, 
the Sarvavajrodaya of Anandagarbha and the Guhyasamajamandalavidhi of 
Diparikarabhadra, the successor of Buddhajnana at Vikramasila, insert ancil- 
lary rites specifically for the averting of danger from the monarch. 223 Moreover, 

to Nalanda and made him the Guardian of the Western Gate of Vikramasila'; p. 
182, 1. 10: bdus kyi ka chen dang po bram ze rin chen rdo rje ni 'The brahmin 
Ratnavajra, the first [occupant of the the position of the] Great Central Pillar of 
Vikramasila'; p. 182, 1. 19: rgyal pos bi kra ma shi la'i *pa (corr. : sa Ed.) tra 
phul 'The king bestowed [on Ratnavajra] the charter of appointment [as the chief 
monk] of Vikramasila' HBI, p. 297 and 301. We may presume that the same ap- 
plied to those who held office as the Gate Guardians of the other three direc- 
tions (Rgya gar chos 'byung, p. 181, 11. 8-10): and to Jnanasrimitra, described 
as the second to hold office at Vikramasila as the Great Central Pillar (p. 183, 1. 
11). King Bheyapala (Abhayapala?), a king otherwise unknown, whom Taranatha 
makes the predecessor of Neyapala (Nayapala [r. c. 1027-1043], the successor of 
Mahipala I), is reported to have bestowed charters of appointment on only sev- 
enty Panditas of Vikramasila (Rgya gar chos 'byung, p. 184,11. 14: bi kra ma 
shi lar ni | pa ndi ta bdun cu tsam gyi *pa tra (corr. : sa tra Ed.) las ma tshugs 
te; HBI, p. 304) Taranatha tells us that for that reason he is not counted among 
the Seven Palas (p. 184, 11. 14-15, HBI, p. 304), that is to say the seven remem- 
bered for their exceptional patronage of the faith. These seven are not listed, but 
Taranatha does say which of the Palas were excluded from the list. The seven 
that remain are Gopala, Devapala, Dharmapala, Mahipala, Mahapala, Neyapala 
(Nayapala), and Ramapala. Other, later appointments recorded by Taranatha 
are those of Dipahkarasrijnana as Upadhyaya at Vikramasila under Bheyapala, 
with responsibility also for Uddandapura (p. 304), the Pramanika Yamari under 
Nayapala (p. 187, 1. 19: bi kra ma shi lar *pa tra (corr. : sa tra Ed.) cher thob 
'He obtained the great charter of Vikramasila'; HBI, p. 308), and Abhayakaragupta 
as Upadhyaya, first at Vajrasana and then at Vikramasila and Nalanda, under 
Ramapala (p. 189, 1. 10-13; HBI, p. 313). I take the term patra here (=patram, 
patrika) to mean an official document bestowing an office and hence by exten- 
sion office or authority bestowed by this means; cf. patrika in Tantralokaviveka, 
vol. 3, p. 191, 11. 3-6, the commentary of the Kashmirian Mahanayaprakasa 
p. 115,8, and Vamakesvarimatavivarana, p. 55 (on the theft of such documents 
by fraudulent Gurus); also the expressions tamrapatram and sasanapatram for 
a royal charter. With the names of Indian Buddhist authors and translators 
we commonly encounter the title Mahapandita (Mkhas pa chen po / Pan chen) 
(also Mahapanditasthavira, Mahapanditacarya, and Mahapanditabhiksu). Among 
Tantric scholars with this title are Atulyavajra, Advayavajra, Abhayakaragupta, 
Anandagarbha, Kuladatta, Darpanacarya, Dipahkarasrijnana, Durjayacandra, 
Naropa, Buddhaguhya, Bhavabhatta, Ratnaraksita, Ratnakarasanti, Ravisrijnana, 
Vaglsvarakirti, Vibhuticandra, Sakyaraksita, and Sridhara. It is perhaps analo- 
gous to the Chinese Buddhist title dashi (Jap. daishi) 'Great Master', which came 
to be bestowed by the Emperor on distinguished monks from the reign of Yizong 
(859-873) onwards; see Forte 1994, pp. 1023-1034. 
223 Anandagarbha, Sarvavajrodaya f. 29rl-2 (a preliminary rite): *manusasthicurna- 


The Saiva Age 

Taranatha relates several occasions on which Buddhist Tantric masters were be- 
lieved to have used Tantric rituals to good effect against the enemies of their 
patrons in times of danger. 224 In some sense, then, these were state monasteries, 
not unlike the great imperial monasteries of Tang China and Japan, 225 rather 

homenasrgvisasahitena (em. [Tib., cited in Ed. mi rus kyi bye ma khrag dang 
dug dang bcas pa dang] : manusasthicurnaho + + + + visasahitena Cod., Ed.) 
mandalavighnam nivaryatmasisyabhupalddisantikahomam kuryat 'After having 
removed [all] impeding spirits from the Mandala by offering into the fire powder 
of human bone mixed with blood and poison he should perform a fire-sacrifice 
for the warding off of dangers from himself, the candidate(s) for initiation, and 
the monarch or other [ruler]'; and Diparikarabhadra, Guhyasamajamandalavidhi 
f. 16vl, vv. 373-374 (a concluding rite): saty eva sambhave tesam pratyekam 
vamapanina \ savyahgusthakam agrhya santim kuryad vidhanatah || trisaptahutim 
ekam va rajho va bhupater atha | dikpalasvatmasantau ca hutva yaceta daksinam 
'With his left hand he should take hold of the right thumb [of the person who has 
been initiated] and make offerings into the sacrificial fire in accordance with the 
prescribed procedure, doing this for each [of the initiates in turn], if that is possible. 
Having made twenty-one oblations or just one to ward off danger from [each of 
these and, then from] the monarch or [lesser] ruler, also from [the Vajracaryas 
who have officiated as] the Guardians of the Directions and himself, he should 
request his fee'. The rite of offering at this point a santikahomah of twenty-one 
oblations for each of the candidates while holding their right thumbs with the 
left hand is derived from Mahavairocanabhisambodhitantra, but the extension 
of that rite in order to protect the king, the Guardians of the Directions, and the 
main officiant himself is an innovation not found there; f. 172v5-6 . . . 173r3-4: 
slob ma sdig dang bral ba kun | de Itar legs par btsud nas ni | de dag zhi bar bya 
ba'i phyir sbyin sreg cho ga bzhin du bya . . . de nas slob ma re re nas | mkhas pas 
lag pa g.yon pa yis | g.yas pa'i mtho bong bzung nas su \ mnyam par bzhag pas 
sbyin sreg bya \ yid ni mnyam par bzhag nas su \ sreg blugs re re las kyang ni | 
gsang sngags cho ga bzhin zlos shing \ nyi shu rtsa gcig sbyin sreg bya | na mah 
sa ma nta bu ddha nam \ om ma ha sha *nti (em. : nta Cod.) ga ta sha nti ka ra 
pra sha ma dha rmma ni rja ta a bha ba sva bha ba dha rmma sa ma ta pra pte 
sva ha \ sbyin sreg rjes la sngags pa yis | slob ma mams la yon bslang ba 'When 
he has in this way introduced all the sin-free disciples [before the Mandala] he 
should duly perform a fire-offering to ward of danger from them. . . . Then the 
learned [officiant], should concentrate himself and make offerings into the fire, 
after grasping the right thumb of each disciple with his left hand. With his mind 
concentrated he should offer twenty-one oblations for each, reciting according to 
the Mantra rite namah samantabuddhanam | om mahasantigata santikara 


After the fire-offering the Mantrin should request his fee from the disciples'. 

224 Rgya gar chos 'byung, p. 178, 11. 4-7; HBI, p. 294 (the Balyacarya of Vikramasila de- 
stroys a Turuska army invading from Bengal); p. 186, 11. 8-11, HBI, p. 306 (Prajna- 
raksita makes offerings to Cakrasamvara when Vikramasila monastery is attacked 
by a Turuska army: the army is struck by lightning, which killed their leader and 
many others, so that they were repelled); p. 197, 1-4, HBI, pp. 326-7 (Lllavajra, 
Tantracarya of Vikramasila, defeats the Turuskas by drawing the Yamaricakra); 
and p. 197, 1. 22-p. 198, 1. 9; HBI p. 328 (Kamalaraksita drives off a Turuska army 
from Vikramasila by throwing enchanted water at them during a Tantric feast 

225 On the imperial Great Monasteries of China and Japan (Ch. ta si, Jpn. daiji [Skt. 


Genesis and Development of Tantrism 

than autonomous, self-governing institutions. 

The Palas' Engagement with Saivism 

The Palas were certainly the most liberal patrons of Buddhist institutions 
in early medieval India, and it was no doubt largely because of this that the 
religion was able to develop and flourish so remarkably in their realm. However, 
it should not be thought that the scale of these rulers' support implies that they 
at least, unlike the other royal patrons of Buddhism that have been reviewed 
here, must have turned their backs on Saivism, starving it of patronage that it 
might otherwise have received. For there is much evidence to the contrary. 

In the ninth century Devapala is praised in a charter of his son Mahen- 
drapala for having built two temples of outstanding beauty during his rule, 
one for the Buddha and the other for the consort of Siva; 226 and Mahen- 
drapala is reported to have established a temple for the emaciated goddess 
Carca (Carcika/Camunda). 227 An eleventh-century Prasasti from Bangarh, 
ancient Kotivarsa in Varendri, also called Devikota and Sonitapura, informs 
us that Nayapala had the Saiddhantika Sarvasiva as his royal preceptor 
(gaudarajaguruh), and that when Sarvasiva retired he passed this office to 
his brother Murtisiva. This implies that Nayapala received Saiva initiation, 
since to initiate the king is fundamental to the Saiva Rajaguru's role. It also 
tells us that at the site of this inscription Mahipala I, Nayapala's predecessor, 
had bestowed a Kailasa-like monastery on Sarvasiva's predecessor Indrasiva. 
Mahipala is described here as a 'knower of reality' (tattvavit), which suggests 
in this Saiva context that he too had received Saiva initiation, which suggests 
in turn that the gift of the monastery was his Guru's daksina. It is probable, 
therefore, that Indrasiva too, like his successors Sarvasiva and Murtisiva, 

mahaviharah]) see Forte and Durt 1984. For Japanese Tantric Buddhist rituals 
of state protection (chingokokka) see May 1967. 

226 gj 42 : 2, 11. 12-13: yo nirmame *sugatasadma grham ca (corr. : sugatasadmagrhan 
ca Ed.) gaurya yat kautukam ca tilakam cajagattraye 'pi. 

227 EI 39:7, the Siyan stone slab inscription of Nayapala, v. 40: mahe[ndra]pdlacarca- 
yd mahendrasadrsodayah | yah sailim vadabhim saile sopdnena sahakarot 'who, 
equal in greatness to Mahendra (Visnu), built for Mahendrapala's Carca a stone 
Vadabhi temple on [her] hill and a flight of steps [leading to it]'. When D.C. Sir- 
car published this inscription he judged that it is probable that the Mahendrapala 
mentioned in this verse is the Gurjara-Pratihara king of that name (EI 39:7, p. 48), 
who ruled c. 885-908. In the light of the discovery of Mahendrapala's Malda in- 
scription (EI 42:2) we may now safely assume that he was the Pala of that name. 
On this goddess see here p. 231. Carcika, Camunda, Carmamunda, and Karnamoti 
are listed as synonymous deity-names in Amarakosa 1.1.46. The name Carcika ap- 
pears in place of Camunda in the Picumata in treatments of the eight Mothers (the 
seven ending with Carcika [Mahesvarl, Brahmam, Vaisnavi, Kaumarl, Vaivasvati, 
Mahendri, Carcika], with Parama/Puram/Aghoresi making up the total). 


The Saiva Age 

had held the office of royal preceptor. 228 I know of no direct evidence that 
Mahipala's successor Vigrahapala III had a Saiddhantika Rajaguru, but it is 
likely that he did, since in his Amgachi copper-plate inscription he is described 
as 'devoted to Siva's worship', 229 and there is evidence which strongly suggests 
that this tradition was still in place under his successor Ramapala. For in the 
twelfth century the South-Indian Saiddhantika Trilocanasiva tells us that his 
preceptorial line descends from a Dharmasambhu (Dharmasiva) who had held 
office as the royal preceptor of "the king of Gauda", a standard expression for the 
Pala rulers. 230 Since three preceptorial generations intervene in that account 
between Dharmasambhu and Trilocanasiva, it is probable that this king was 
Ramapala. 231 

228 The Bangarh Prasasti of Murtisiva (Sircar 1983b), found at Sivavati (mod. 
Sibbadi) in the vicinity of Kotivarsa, 11. 8-9: 9 srlman indrasivah sphutam ha- 
riharaprayam sivendrakrtim bibhrad vamsavibhusanam samabhavac chisyo 'sya 
punyatmanah | yasmai kancanapuhjamahjuracitaprasadamerusphuratkailasabha- 
matham daddv iha mahlpalo nrpas tattvavit; 11. 11-12, reporting that Indrasiva's 
successor Sarvasiva was the royal preceptor of Nayapala: rajho srlnayapalasya gu- 
rus tattvavidam varah \ srlman sarvasivas tasya sisyo 'bhud bhusanam bhuvah; 
and 11. 13-14, reporting that Sarvasiva resigned his office as the Gaudarajaguru in 
favour of his brother Murtisiva: 14 yenavarjitagaudarajagurutalaksmlr nijabhra- 
tari srlman milrtisive nivesya vipinavdsam svayam vanchata | kslrodarnavavama- 
nthanotthitamilallaksmlm svasisye harav aropyaharato visam pasupater vrttantam 

229 EI 15:18, 11. 17-19 (v. 12): plta<h> sajjanalocanaih smararipoh pujanuraktah 
sada samgrame caturo 'dhikas ca haritah kalah kule vidvisam \ caturvarnyasama- 
srayah sitayasahpunjair jagad rahjayan srlmadvigrahapaladevanrpatir jajne tato 
dhamabhrt 'From [Nayapala] was born the illustrious king Vigrahapaladeva, who 
was drunk by the eyes of the virtuous, ever devoted to the worship of Siva, more 
skilled in battle than Indra, the god of Death to the families of his foes, support of 
the four caste-classes, white-washing the world with the multitudes of his stuccoed 

230 See, e.g., in a pedestal inscription of the reign of Palapala (r. c. 1165-1199): srlgau- 
desvarapalapalapadanam (HUNTINGTON 1984, p. 239, no. 59) and the Sarnath in- 
scription of Mahipala (HULTZSCH 1885), v. 2: gaudadhipo mahlpalah. 

231 Colophonic verses at the end of Trilocanasiva's Somasambhupaddhativydkhyd 
(IFP, MS Transcripts 457 [Tl] and 170 [T2]; edition in Brunner 1963- 
1998, Pt. 4, pp. 422-427 [B]): 1 srlcedirajabhuvi *saivajanakarakhyasrlgolakl- 
yamathabhavasivas ca yo 'sau (saivajanakarakhya T2B:saivajanakarakhyas T2 
• srlgolaklyamatha conj. : srlkolaklvimala T1T2 : srlgolaklvimala B • bhavasivas 
ca yo 'sau conj. :bhavasivasayosau T1T2B) | tadvamsajah sivamatagamalaksa- 
vetta srldharmasambhur iti gaudapatindranathah \\ 2 tasmad asav anala- 
sankaradesiko 'bhud divyagamambunidhir *lhitakalpavrksah (Tl : itikalpavr- 
ksah B) | svargaukasam api padam vacasa labhante *yasyaiva (conj. B -.yasyaika 
T1T2) janmamaranaika*bhayan (T2 :bhayamTl) nirastah \\ 3 *srlgolaklyasam- 
tanavyomavyapl (golaklya T2B :kolaklya Tl) tatah sivah \ srisomasambhur ity 
aslt kalau lokahitaya vai \\ 4 jnanasaktivapus tasmaj jnanasambhuh sada- 
sivah | yenedam dyotitam sarvam saivajhanamalarcisa || 5 somarkavamsanrpa- 
mauli*vilolitahghrir (T2B -.vilolitahghri Tl) vidvajjanananasarojadivakaro mam 


Genesis and Development of Tantrism 

There is other evidence of these kings' engagement with Saivism. The poet 
Samdhyakaranandin describes king Madanapala, Ramapala's second son, as a 
devotee of Siva; 232 and a pedestal inscription of 1026 recording renovations of 
Buddhist structures at Sarnath by two Pala princes Sthirapala and Vasantapala, 
also tells us that Mahipala I had engaged them to have hundreds of temples of 
Siva, Citraghanta, and other deities built in Benares and that he did so after 
offering obeisance at the feet of the Guru Vamarasi of that city, who, as we can 
infer from his name in -rasi, was a Saiva ascetic of the Atimarga. 233 

| dlndndhasurikrpandtithi*pdrijdtah (corr. -.pdrijdta T2B warajata Tl) srljhdna- 
sambhur anisam malinam pundtu '[1] In the land of the king of Cedi [livedl 
Dharmasambhu, a spiritual descendant in the lineage of the famous Bhavasiva 
[=Sadbhavasiva/Prabhavasiva, founder] of the venerable monastery at Golaki. He 
mastered one hundred thousand [verses] of the scriptures of the religion of Siva 
and became the Lord [Guru] of the King of Gauda. [2] His successor was the fa- 
mous teacher Analasiva, an ocean of the celestial scriptures, a tree of paradise 
that granted every wish, one through whose instruction men attained the world 
of the gods, free of the unique terror of birth and death. [3] His successor was 
Somasambhu, a Siva who for the good of mankind [was the sun whose light] filled 
the sky of the venerable lineage of Golaki. [4] His successor in [this] Kali age was 
Jnanasambhu, the very embodiment of [Siva's] power of knowledge, [a] Sadasiva 
who illuminated this universe with the pure radiance of his understanding of Siva's 
teachings'. [5] His feet were caressed by the crowns of kings of the lineages of both 
the moon and the sun. He was a sun to the lotuses that are the faces of the learned. 
He was the tree of paradise to the needy, to the blind, to scholars, to the wretched, 
and to uninvited guests. May Jnanasambhu ever [continue to] cleanse me [as his 
disciple], impure as I am'. The king of Cedi referred to at the beginning of this 
passage is the Kalacuri and his land is Dahaladesa, the region of central India ap- 
proximately comprising within modern Madhya Pradesh the Jabalpur District, and 
parts of the Satna, Panna, and Rewa Districts. 

232 Rdmacarita 4.35b: sivapranayi. 

233 rp^g gfjmath inscription of Mahipala (HULTZSCH 1885): om namo buddhdya \ *va- 
rdnasisarasydm (corr. -.varanasisarasyam Ep.) guravasrlvdmardsipdddbjam | drd- 
dhya namitabhupatisiroruhaih saival*ddhisam (?) || isdnacitraghantddikirtiratna- 
satdni yau | gaudddhipo mahlpdlah kdsydm srimdn akdrafyat] || saphalikrtapdndi- 
tyau bodhdv avinivartinau \ tau dharmmardjikdm sdhgam dharmmacakram pu- 
nar nnavam \\ krtavantau ca navindm astamahdsthdnasailagandhakutim | etdm 
sristhirapdlo vasantapdlo 'nujah srimdn 'Obeisance to the Buddha. Sthirapala 
and his younger brother Vasantapala, whom the Glorious Mahipala, the ruler of 
Gauda, caused to erect hundreds of fine temples for Siva, Citraghanta, and [other] 
gods in Kasi after worshipping the venerable Gurava Vamarasi's feet, the lotuses 
that beautify the lake that is Varanasi, with [strands of] duckweed *clinging to 
them (?) in the form of the hair of the kings that bow down to them, have made 
the Dharmarajika, a new Dharmacakra together with its ancillaries, and a new 
Buddha-shrine from stones of the eight sacred places, having made their learn- 
ing bear fruit, refusing to turn back in their quest for enlightenment'. The read- 
ing saivalddhlsam is surely a mistake, for if it were sound it could only yield 
the absurd meaning 'overlord of duckweed'. The meaning required by the con- 
text would be secured by saivaldsahgam. This has the advantage that it echoes 
a verse in Kalidasa's Kumdrasambhava (5.9), which is likely to have been in the 


The Saiva Age 

Similarly, the Bhagalpur copper-plate inscription of Narayanapala (r. c. 860- 
917) records his establishing a Siva and granting a village to it and the associ- 
ation of Pasupatacaryas (pasupatacaryaparisat) attached to the foundation; and 
though it gives him the epithet paramasaugatah it reports that he had been re- 
sponsible for the building of a vast number of other temples for this deity 234 

We have even more striking evidence of this kind in the case of Nayapala. 
His Siyan stone slab inscription {EI 39:7) devotes most of its sixty-five verses 
(21-63) to detailing an extensive program of royal temple building and image 
installation undertaken throughout the Pala realm. Damage to the inscription 
has removed the name of the king who was responsible for this program, but it 
is extremely unlikely that it was other than Nayapala, since the account follows 
immediately on that of his martial exploits, following those of his predecessors. 
These pious activities comprise the construction of a temple topped by golden 
lions and a finial, evidently therefore a Vadabhi temple for a goddess, 235 with 
a temple of Siva and an attached two-storied monastery (matho dvibhumih) for 
the accommodation of ascetics to its south (v. 24), a temple with a [golden] finial, 

memory of the author of the inscription, to the effect that during the austerities 
that Parvati undertook to win the hand of Siva her face was just as charming with 
her ascetic's braids as it had been with her elegantly adorned coiffure; for, says 
Kalidasa: "The lotus is not beautiful only when when lines of bees hover about it 
but even when [strands of] duckweed cling to it" (na satpadasrenibhir eva pahkajam 
sasaivalasangam api prakdsate). However, this solution has the weakness that 
it is not open to any obvious explanation of how the error arose. Perhaps the per- 
son who wrote the letters on the stone before they were engraved was thinking 
of Vamarasi's official status in Benares. If that, as is very likely, was as the ab- 
bot of a Saiva monastery, then the error -adhlsam might be the result of the in- 
trusion into his mind of an expression such as saivddhlsam, saivamathddhisam, 
or saivalay ■adhlsam. For the expression mathadhisah (=mathddhipatih) see, e.g., 
Rdjatarahgini 7.298ab: bhattdrakamathadhisah sddhur vyomasivo jati; and the 
anonymous Kumdrapdladevacarita v. 51a: tarn nimantrya mathadhisarn (called 
mathddhipatih in v. 49b). But this would be more convincing if the reading 
corrupted were closer to saivalddhlsam in written appearance or pronunciation. 
Citraghanta has her temple in Benares near that of Siva Citraguptesvara as one of 
the Nine Durgas. The sense intended may be that he had [new] shrines built for all 
nine of these goddesses. 
234 HULTZSCH 1886, 11.28-29: paramasaugato mahdrdjddhirdjasrwigrahapdla- 
devapdddnudhydtah paramesvarah paramabhattdrako mahdrdjddhirdjah 
srimanndrdyanapdladevah ...; 11.38—41: matam astu bhavatdm \ kalasapote 
mahdrdjadhirdjasrindrdyanapdladevena svayamkdritasahasrdyatanasya tatra 
pratisthdpitasya bhagavatah sivabhattdrakasya pdsupata-dcdryaparisadas ca | 
yathdrham pujdbalicarusattranavakarmddyartham sayandsanagldnapratyaya- 
bhaisajyapariskdrddyartham | anyesdm api svdbhimatdndm | svaparikalpita- 
vibhdgena anavadyabhogdrtham ca \ yathoparilikhitamuktikdgrdmah .... I agree 
with HULTZSCH that svayamkdritasahasrdyatanasya here means '[Siva] for whom 
he [Narayanapala] himself has built a thousand temples'. 


Genesis and Development of Tantrism 

presumably for Siva, since it was equipped with eleven [subsidiary] shrines in 
which the eleven Rudras were installed (v. 25), a Vadabhi temple for the Mother 
Goddess 236 and a series of temples for the Nine Durgas, 237 a lofty temple for 
Siva Hetukesvara at Devikota, 238 a temple of Siva Ksemesvara with a golden 

235 Verse 23ab: [sujdhdsubhram kdncanasimhakumbhasirasam That a Vadabhi 

temple housing an image of a goddess should be distinguished from others by being 
surmounted by [twol lions and a finial, and that Vadabhi temples are principally for 
the housing of goddesses, is prescribed in the Saiva Pratisthatantras, Tantras, that 
is, which specialize in temple construction and installation. See Mayasamgraha, 
f. 28r-v (5.86c-89): vasvamse sodasatydgdt suryasamvardhitdyatih || 87 catur- 
danat purah siddhasukdghro vadabhih smrtah \ prdsddo vyaktalihgesu netaresu- 
dito budhaih || 88 vistdrdd dvigunotsedhahphamsddikrtasamvrtih \parsve simha- 
dvayopeto madhye kalasabhusitah || 89 padaikasdrdhabhittir vd sapdda- 
dvigunonnatih \\ visesato 'mbikddlndm samnidhisthdnam iritam; ibid., f. 29v 
(vv. 119-121): vadabhyam ambikadevyah kesarl garudo hareh \ sriyo dvipo 
vrsah sambhoh savituh kamalo 'thavd \\ tad anyesdm ca devdndm svdyudham 
vd hitam param | svacihnaparamam yad vd nijakalpoktam eva vd || yad utpatti- 
sthitidhvamsakdranam visvatomukham | bhdti sarvdtmano murdhni sd cudd ga- 
ditd budhaih; f. 28v (5.89cd), referring to the Vadabhi type of temple: visesato 
'mbikddlndm samnidhisthdnam Iritam. The sections of this and other unpublished 
Saiva works (Brhatkdlottara, Pihgaldmata, Devydmata, and Mohacudottara) that 
deal with the building and design of the various kinds of temple are being edited, 
translated, and analyzed in a doctoral thesis being prepared by my pupil Elizabeth 

236 y erge 26: mdtuh krte 'traiva *suvarnakumbhabhrdjisnumurdhdm (em. : suvarna- 
kumbhabhardjisnumurdhdm Ed.) valabhim sildbhih \ [20 syllables obliterated! 

237 Verse 27: saildni mandirdny atra mandardnkdni ydni ca\ + + + + + + + + 
krtd yd nava candikdh 'and here stone temples of the Mandara kind . . . the 
Nine Candikas'. The Nine Candikas are surely the eighteen-armed form of 
MahisasuramardinI Durga known as Ugracanda and her eight sixteen-armed 
ancillaries Rudracanda, Pracanda, Candogra, Candanayika, Canda, Candavati, 
Candarupa, and Aticandika. They are nine to match the nine days of the autum- 
nal Navaratra festival. For these goddesses, also called the Nine Durgas, see Ag- 
nipurdna 50.7-11 and 185.3-10; and Vidyapati, Durgdbhaktitarahgini, p. 198. That 
Nayapala had [nine] temples built for these goddesses is in keeping with the pre- 
ferred option of Agnipurdna 185. 3cd: durga tu navagehasthd ekdgdrasthitdthavd 
'Durga may be in nine temples or one'. For a Paddhati for the worship of Ugracanda 
and her ancillaries see Ugracanddprakarana. 

238 Verse 28ab: devikote hetukesasya sambhor yah prdsddam sailam uccair akdrsit. 
For the Hetukesvara of Devikota/Kotivarsa (modern Bangarh) see Sander- 
son 2001, fn. 4, p. 7; also Picumata f. 8r3-4 (3.119c-123), which requires 
the installation of Hetukesvara as Bhairava in the northeastern segment of 
the initiation Mandala: isdne tu disdbhdge kotivarsam. prakalpayet || 120 
vatam tatra samdlikhya tatra sulodakam likhet \ diksu caiva vidiksu ca 
sulaprotd likhet tathd || 121 sula tasydgrato likhya kundasyaiva mahdtape 

| pattisam purvato nyasya vatasyddhas tato priye || 122 astapatram likhet 
padmam tathaiveha na samsayah \ hetukesvaram dlikhya saddsivatanus tathd 
|| 123 karnikdydm mahddevi mahdbhairavarupinam | rudrdstakasamopetam 
purvavad devi cdlikhet; and Nisisamcdra f. 17v (4.20—21): kotlvarse karnamoti 


The Saiva Age 

finial and a water reservoir, 239 a temple of Siva Varaksesvara together with a 
monastery and reservoir, 240 a temple of Visnu (v. 33), a temple of Ghantisa and 
of Bhairava surrounded by the sixty-four Mothers 'in his own city', 241 a temple of 
Siva Vatesvara at Campa, 242 and a Vadabhi temple on a hill-top with a flight of 

mahdbalakulodbhavd | sulahastd sthitd devi sarvayogesvaresvarl || tasmin ksetre 
sthitd devi vatavrksasamdsritd \ ksetrapdlo mahdkdfyo] hetuko ndma ndmatah. 
The origin myth of the cult of Hetukesvara, Bahumamsa (=Karnamoti/Camu- 
nda/Carcika), and the other Mothers (Matrs) at Kotivarsa is narrated in chapter 
171 of the early Skandapurdna-Ambikdkhanda. Siva promises the Mothers there 
that he will compose Tantras of the Mothers (mdtrtantrdni) to guide their worship. 
The names of these reveal them to be the Yamalatantras; see Sanderson 2001, 
pp. 6-7, fn. 4. 

239 Verse 30: ksemesvarasydyatanam ksemahkaro grdvamayam smardreh | cakdra 
yo murdhni diptdyatasdtakumbhakumbham vyadhdt tatra mahdsaras ca. In 
a passage describing Varendri (3.1-27) in the Rdmacarita, completed in the 
time of Madanapala (r. c. 1143-1161) but relating events that occurred dur- 
ing the reign of Ramapala (r. c. 1072-1126), Samdhyakaranandin devotes 
six verses to the deities of the region (3.2-7). There Ksemesvara appears 
with Hetvisvara or with Hetvisvara and Candesvara as one of only two or 
three deities individualized by a personal name (3.2-5: kurvadbhih sam devena 
srihetvisv arena devena \ candesvardbhidhdnena kila ksemesvarena ca sandthaih 
|| . . . sambhdvitdkalusabhdvdm), the others mentioned being generic: the twelve 
Adityas, the eleven Rudras, Skanda, Vinayaka, the Vasus, the Visvadevas, and the 
Lokapalas. Hetvisvara here is surely identical with the Hetukesvara of Kotivarsa 
mentioned above. As a synonymous form it was probably substituted for metrical 
convenience. It is not clear from the Sanskrit whether Samdhyakaranandin in- 
tended Candesvara to be understood as an alias of Ksemesvara or as the name of 
third local Siva. I am not aware at present of any external evidence that removes 
this doubt. 

240 Verse 32: . . . matham ca sarasim ca | dhdma varaksesvara iti sambhor api sailam 

241 Verse 35: ghantisam yah svanagare nyadhdt ksemdya dehindm | catuhsastyd 
ca mdtrndm paritam tatra bhairavam. This Ghantisa is perhaps a double of 
the Mahaghantesvara/Mahaghanta identified by the Picumata (3.77c-83) as the 
Bhairava of Viraja, modern Jajpur in the Cuttack District of Orissa, formerly 
the capital of the Bhauma-Kara kings: dgneye (em. : dgneyam Cod.) virajayam 
tu trikutam tatra cdlikhet | 78 ndndvrksasamdkirnam ulukais copasobhitam 

| nandin ca chagalam caiva kumbhakarnam mahdbalam || 79 hetukam tatra 
devesam smasdnena *samam nyaset (conj : samabhyaset Cod.) | tatropari likhec 
chaktim karanjam ca mahddrumam || 80 tasyddhastdl likhet padmam astapatram 
sakarnikam | karnikdydm likhed devam mahaghantam tu bhairavam 
|| 81 katidese tathd caiva ghantdsaptavibhusitam \ rudrdstakasamopetam 
bhairavdkdrarupibhih || 82 rudrdndm bdhyato devi yoginyah sat samdlikhet | 
yamaghantd kardld ca mahdjihvd khardnand || 83 kardli danturd caiva ndmais 
caitdh prakirtitdh | rudracakram ca samvestya saddiksu ca kramdt sthitdh; and 
3.136cd (f. 8v2-3): dgneye mahaghantesvaram likhet; 30.25cd: dgneyapahkaje 
caiva mahaghantesvaram nyaset. Ghantisa- is evidently Ghantesa- modified by 
Middle-Indie Sandhi (-aid + i- > -'!-). 

242 Verse 38: vatesvarasya vikatas campdydm dlayo 'smabhih \yena vyadhdyi navamah 
kuldcala ivocchritah. Campa was the capital of Ahga in the eastern part of the 


Genesis and Development of Tantrism 

steps for the emaciated goddess Carca (Carcika) previously established by king 
Mahendrapala, 243 the re-excavation of the step-well (vapi) of the sage Matahga 
at Dharmaranya, the building of a lofty temple of Siva Matangesvara at that 
site (v. 43), 244 the building of a temple of Laksmi (v. 44), the erecting of a golden 
Trisula at Sagara (v. 45), 245 the building of a temple of the Sun-god (v. 46), the 
provision of a golden cover for [the Liriga of] Siva Vaidyanatha, 246 the installa- 
tion of a golden finial on the temple of Siva Attahasa (v. 50), 247 the making of 
a silver image of Sadasiva, golden images of Candika and Ganesa (v. 53) with 
golden pedestals, a Moon-god, a Sun-god of silver, a golden lotus engraved with 
images of the Nine Planets (vv 54-55) — all these are ancillary deities of Saiva 
worship — , and a bejewelled golden Siva (v. 56), the building of a monastery and 
the installation in it of an image of Visnu in his [Pancaratrika] Vaikuntha form 
(v. 61), and the building of a high Vadabhi temple for the goddess Pirigalarya. 248 
A few other temples and one monastery are mentioned in the inscription (vv 21- 
22, 31, 36-37, 39, 41-42, 47, 52, and 59-60), but their affiliation is not stated or 
has been lost through damage to the stone. 249 

It is striking that most of these constructions and images are Saiva or Sakta 
Saiva and that not one is Buddhist. It is unlikely, however, that Nayapala had 
rejected the Buddhist leanings so marked in this dynasty. For in addition to the 
evidence of his being called paramasaugatah there is the fact that Taranatha 

modern state of Bihar. 

243 Verse 40: mahendrapdlacarcdyd mahendrasadrsodayah | yah saillm vadabhlm 
saile sopdnena sahdkarot. Carca/Carcika is the fearsome emaciated goddess com- 
monly known as Camunda or Karnamoti; see here p. 231. 

244 Dharmaranya is at Gaya in southern Bihar. Its Matariga hermitage, its step-well of 
Matariga, and its temple of Matangesvara are mentioned in Agnipurdna 115.34-36. 

245 This is probably Gahgasagara/Gangasagarasamgama, where the Ganges flows into 
the Bay of Bengal, listed in Saiva sources as one of the Saiva sacred power sites, 
e.g., in the list of the siddhiksetrdni given in the Nisvdsatattvasamhitd, f. 42rl— 3 
(Nisvdsaguhyasutra 1.29-33b). 

246 y erge 4g. kholam akdri rukmaracitam srivaidyandthasya tat. Temples of Siva 
Vaidyanatha are found in various parts of the subcontinent. However, SlRCAR is 
no doubt correct in his annotation of this inscription (EI 39, p. 41) that this is the 
Vaidyanatha of Deoghar (24°29' N, 86°42' E ) in Jharkhand, this being revered as 
one of Siva's twelve Jyotirlingas. 

247 Perhaps at Attahasa, now Labpur (23°50' N, 87°49' E) in the Bhirbhum Dis- 
trict of Bengal. The name of the Siva at this Saiva and Sakta sacred site 
is Mahanada (e.g. Matangapdramesvara, Vidydpdda 20.53ab: mahdnddasya 
ndthasya cdttahdsakhyam eva hi \ vimalam vimalasyoktam sthdnam rudrasya 
sobhanam); but Attahasa being nearly a synonym as well as the name of the site 
may have been an alias. 

248 Verse 63cd: iyam api valabhl grdvabhir uttuhgd pihgaldrydydh . 

249 In addition v. 34 records the founding of a hospital (drogyasdld), and v. 57 gifts to 


The Saiva Age 

reports that Nayapala had a Buddhist preceptor in the person of Mahavajrasana 
Punyakaragupta. 250 

Buddhist Kings of Eastern India and their Commitment to Brahmanism 

Nor is it the case that royal devotion to the Buddha in eastern India dur- 
ing this period weakened in this region the traditional commitment of Indian 
rulers to the imposition and preservation of the caste-based brahmanical social 
order in which Saivism was embedded. For in the Neulpur grant of the Bhauma- 
Kara king Subhakara I his grandfather Ksemarikara is described both as a Bud- 
dhist and as having ensured that the members of the caste-classes and disci- 
plines observed their prescribed roles; 251 in his Terundia copper-plate inscription 
Subhakara II, the grandson of Subhakara I, is given the epithet paramasaugatah 
yet is also commended for having 'propagated the system of uncommingled caste- 
classes and disciplines proper to the [perfect] Krta Age following the unexcelled 
[brahmanical] scriptures'; 252 the Pala Dharmapala is described in a grant of his 
son Devapala both as a paramasaugatah and as taking measures to ensure that 
castes that erred were made to adhere to their respective duties, thereby dis- 
charging his father's debt to his deceased ancestors; 253 and Vigrahapala III is 

250 Rgya gar chos 'byung, p. 185, 11. 7-9: rgyal po 'dis rdo rje gdan pa chen por grags 
pa la mchod de | de dge bsnyen gyi dus kyi mtshan pu nya shri \ rob tu byung ba'i 
mtshan pu nya a ka ragu pta'o 'This king [Neyapala] venerated [the teacher] called 
Mahavajrasana. During his time as a lay Buddhist, his name was Punyasri. His 
ordination name was Punyakaragupta'; HBI, p. 305. In Taranatha's text the name 
of the king is given as Neyapala. But there can be no doubt that it is Nayapala 
that is meant. For there is no other Pala whose name 'Neyapala' approximates, and 
Taranatha's chronology of Neyapala fits this king's reign. He relates that his reign 
began shortly before Dlparikarasrljfiana (Atlsa) left for Tibet, which is not far out, 
since Nayapala came to the throne in approximately 1027 and Dlpankarasrljiiana 
set out for Tibet in 1038. 

251 EI 15:1, 1.2: svadharmaropitavarnasramah paramopasako . . . srlksemahkara- 

252 EI 28:36, 11. 8-10: paramasaugatafh] . . . niratisayasastranusarapravartitakrtayu- 

253 The Mungir copper-plate grant of Devapala, KlELHORN 1892, p. 255, 1.28: 
paladevapadanudhyatah paramasaugatah paramesvara<h> paramabhattarako 
maharajadhiraja<h> srlman devapaladeva<h>; and 11.8-9 (v. 5): sastrarthabhaja 
calato 'nusasya varnan pratisthapayata svadharme \ sridharmapalena sutena so 
'bhut svargasthitanam anrnah pitrnam '[Gopala] became free of his debt to his 
ancestors in heaven through his son Dharmapala, who, adhering to the teachings 
of the [brahmanical] Sastras, after chastising those [members of] caste-classes 
that stray makes them adhere to their prescribed duties'. Cf. Visnudharmottara 
2.65.55: varnasramavyavasiha tu tatha karya visesatah | svadharmapracyutan 
raja svadharme viniyojayet 'The king must above all establish the castes-classes 
and disciplines with the proper distinctions between each. He should force those 


Genesis and Development of Tantrism 

described in his Amgachi copper-plate as the support of the four caste-classes. 254 
Moreover, most of the surviving inscriptions of the Palas, Candras, and Bhauma- 
Karas record grants which they made in favour of Brahmins. The Rampal copper- 
plate grant of the Candra Sricandra strikingly exhibits the extent to which this 
double allegiance was unproblematic for such Buddhist donors. Following a prac- 
tice widely attested in non -Buddhist donative inscriptions the gift of land is said 
to have been made over to its brahmin recipient after the pouring of water and 
the performance of a fire-sacrifice, in this case a kotihomah. 255 This is simply 
adapted to the donor's faith by dedicating the offerings to the Buddha rather 
than to Siva or Visnu. 256 

It seems, then, that royal patronage, reflecting no doubt the balance of alle- 
giance in the wider population, ensured that Buddhism, for all the liberal sup- 
port it received from the Palas, was in no position to oust or diminish Saivism, 
even in this region. The monasteries themselves reflect this symbiosis. The 
excavations at Somapura revealed an abundance of non-Buddhist deities, par- 
ticularly Siva, among the stone relief sculptures around the base of the central 
temple and the very numerous terracotta plaques that decorated its walls. 257 

who fall from their prescribed duties to carry them out'; and the Bhagalpur plate of 
Narayanapala, HULTZSCH 1886, v. 2cd: maryadaparipalanaikaniratah sauryalayo 
'srndd abhud dugdhambhodhivilasahasamahima sridharmapalo nrpah 'After 
him came King Dharmapala. He was solely dedicated to the maintenance of the 
boundaries [between the caste-classes and disciplines]; he was the very abode of 
heroism [in warl; and the glory [of his fame] shone dazzlingly white like the ocean 
of milk '. 

254 EI 15:18, v. 13c: caturvarnyasamasrayah. 

2SB On the brahmanical kotihomah see Sanderson 2005a, pp. 382-383. 

256 EI 12:18, 11.28-29: vidhivad udakapurvakam krtva kotihomam bhagavate 
bhagavantam buddhabhattarakam uddisya mdtdpitror atmanas ca punyayasobhi- 
vrddhaye . . . 'According to rule, after pouring water [upon the hand of the re- 
cipient] and after performing a kotihomah for the Lord and dedicating it to the 
Lord Buddha, to add to the merit and fame of my parents and myself . . . '. Cf., 
e.g., EI 21:37 (the Saktipur copper-plate of Laksmanasena, r. 1179-1206), lines 
42-44: vidhivad udakapurvakam bhagavantam srlnarayanabhattarakam uddisya 
mdtdpitror atmanas ca punyayasobhivrddhaye; EI 21:28 (the Palanpur plates of 
Caulukya Bhimadeva of Gujarat), A.D. 1063, 11. 5—6: mahesvaram abhyarcya mdtd- 
pitror atmanas ca punyayasobhivrddhaye .... We find a similar case in the Amgachi 
grant of Vigrahapala III (EI 15:18, 11. 35-40), but with the omission of the fire- 
sacrifice: mdtdpitror atmanas ca punyayasobhivrddhaye bhagavantam buddha- 
bhattarakam uddisya .... 

257 Dikshit 1938, pp. 39, 41-42, 49, 50, 52, 53, 54, and 58, commenting (p. 58) that 
brahmanical and Buddhist gods are equally and promiscuously represented on the 
terracotta plaques, and that among the brahmanical deities Siva is the most fre- 
quently represented both on those and in the stone relief sculptures. For the forms 
of Siva found here see his Plates XXXa-d, XXXIa-e, XXXIX/ 1 (Linga), XLI d-2, and 
XLIV a and e, LVIe (Mukhalinga), and LVIIIa (Umamahesvara). 


The Saiva Age 

Excavations of the Vikramasila monastery also uncovered a mix of Buddhist 
and predominantly Saiva non-Buddhist images, the latter Siva, Umamahesvara, 
Siva and Parvati, Bhairava, Mahisasuramardini, Parvati, Kaumari, Camunda, 
Ganesa, Kartikeya, the Navagraha, Vrsabha, Visnu, and Surya. 258 

Joint Patronage of Buddhism and Saivism in the Kingdoms of the Khmers, 
Chams, and Javanese 

Much the same phenomenon can be seen in Southeast Asia among the 
Khmers, the Chams, and the Javanese. Among the first the dominant religion 
was Saivism until the rise of the Theravada that accompanied the decline of 
Angkor, and Tantric Buddhism, even when it enjoyed short periods of promi- 
nence through exceptionally determined royal patronage, found itself bound, as 
I have shown elsewhere, to accommodate its rival. 259 

In the kingdoms of the Chams, speakers of an Austronesian language who 
inhabitated the plains along the coast of the South China Sea in what is now the 
central part of Vietnam, most of the inscriptions that have survived, in Sanskrit 
and Old Cham, ranging in time from the fourth to the fifteenth centuries, record 
acts of royal piety to Siva or to goddesses identified with his consort. There 
are also a few from the ninth and tenth centuries that record the installation 
of Tantric Mahayanist Lokesvaras, the construction of associated Viharas, and 
land-grants to these. But as in eastern India we find in these that single donors 
supported both religions. Indeed the situation is more striking here because in 
all but one case each of these inscriptions records a person's practising both kinds 
of patronage, Buddhist and Saiva. 260 Thus in the Bakul stele of 829 a Buddhist 
monk Sthavira Buddhanirvana records that his father Samanta has donated two 
Viharas to the Buddha and two temples to Siva. 261 The Dong Duong stele of 875 
records that King Jayendavarman alias Laksmlndra enshrined a Laksmindra- 
lokesvara and an associated Vihara, yet the bulk of this long inscription is de- 
voted to the praise of the Siva Bhadresvara, who, we are told, is the source of this 
dynasty's power and prosperity 262 The Nhan-bieu stele records that in 908 Pov 

258 JAR 1974-75, p. 7; 1975-76, p. 7; 1976-77, p. 11; 1977-78, p. 15; 1978-79, p. 43; 
and 1979-80, p. 13. 

259 On the co-existence of Saivism and Tantric Buddhism in the Khmer kingdom of 
Angkor see Sanderson 2005a, pp. pp. 421-435. 

260 The exception is the An-thai stele of 902 (Huber 1911, pp. 277-282), which records 
that the Buddhist monk Sthavira Nagapuspa, a close associate of Bhadravarman 
II, installed a Pramuditalokesvara, and also that this king made a land-grant to the 
associated monastery (Pramuditalokesvaravihara). 

261 ISCC, pp. 237-241. 

262 p INOT 1904a, pp. 84-99. 


Genesis and Development of Tantrism 

klun Sudanda[va]sa and his son Pov klun Dharmapatha installed a Siva Deva- 
lirigesvara and in 911, the year of the inscription, built a Vihara for a Vrddha- 
lokesvara, which is to say, a Vihara associated with a deity Vrddhalokesvara, 
which was installed there around this time since it is evident from its name that 
it was established with the name of their grandmother, princess Lyari Vrddha- 
kula, the grandmother of the senior wife of Jayasimhavarman I. 263 A stele at 
Mi-son of 1092 records that King Jayendravarman (alias Paramabuddhaloka), 
described as versed both in the Mahayana and in the brahmanical Dharma- 
sastras, established two Buddhist deities, a Buddhalokesvara and a Jayendra- 
lokesvara, but also two goddesses, a Jayendresvari, and an Indragaurisvari, both 
probably Saiva, and between 1085 and the year of the inscription gave to Siva 
Isanabhadresvara a Liriga-sheath of gold and silver alloy adorned with jewels, 
an inner shrine of sandalwood, silver, gold, and jewels, various items of gold and 
silver, elephants, and male and female slaves, and beautified his temple with 
silver and gilded its pinnacles. 264 

This co-ordination of the two faiths is also evident in eastern Java. The 'Cal- 
cutta' stone inscription of Airlangga (c. 1010-1050), founder of the kingdom of 
Kahuripan, reports in its Old Javanese section that he was consecrated as the 
king in 1019/20 by Buddhist (Saugata), Saiva (Mahesvara), and Mahabrahmana 
dignitaries; 265 and much evidence of the simultaneous royal support of both 
Saivism and Buddhism during the Singhasari and Majapahit periods (1222- 
1292, 1293-c. 1500) is present in the Old Javanese poem Nagarakrtagama, also 
called Desawarnana, completed in 1365 by Mpu Prapanca during the reign of 
Hayam Wuruk of Majapahit, consecrated as Rajasanagara (1350-1389). We 
learn from this work that both Saiva and Buddhist priests participated in pe- 
riodic ceremonies for the benefit of the realm within the great courtyard in- 
side the royal gate of the palace compound, 266 that the administrative heads 

263 Huber 1911, pp. 299-311. 

264 FlN0T 1904b, pp. 970-975 


de Casparis 1992, pp. 482-483; Kern 1885 and 1913, p. 104, 11.14-15: matah 
yan rake halu sri lokesvaradharmmawamsa airlahgdnantawikramottuhgadewa- 
sangjnd kdstwan sri mahdrdja, de mpuhku sogata mahesvara mahabrahmana iri- 
kang sdkakdla 941 'Wherefore he was confirmed with blessings by the high digni- 
taries of the Buddhists, Saivas, and Mahabrahmanas under the name of Rake Halu 
Lokesvara Dharmavamsa Airlangga Anantavikramotturigadeva in Saka 941'. 
266 Nagarakrtagama 8.3-4; PlGEAUD 1960-1963, vol. 4, p. 13. This event is referred 
to by PlGEAUD in his translation (1960-63, vol. 3, p. 10) as "purification (cere- 
monies)". The term used here is the Sanskrit prdyascittam (8.3d: prdyascitta ri 
kdlaning *srawana [conj. PlGEAUD : grahana Cod.] phalguna makaphala hay- 
waning sabhuwana). The function of the ceremony, therefore, was expiatory: to 
cancel the effects of any errors, omissions, or excesses in observances and con- 
duct during the period since the last performance. Kern, accepting the reading 


The Saiva Age 

(dharmadhyaksa) of these two communities had official quarters in the east and 
west to the south of the royal compound, 267 and that his sovereign was dedicated 

grahana phalguna, took the occasion to be an eclipse during the month Phalguna. 
As PlGEAUD saw, this is implausible. He therefore proposed that grahana is an 
error for srawana 'the month Sravana', making this ceremony bi-annual and not- 
ing that the resulting timing coincides with that of the two major festivals of the 
Majapahit court (1960-63, vol. 2, pp. 21-22). A trace of this co-functionality has 
survived into modern times on the island of Bali, where there are both Saiva and 
Buddhist priests (padanda), with the latter now forming a small minority, about 1 
in 10 and less than twenty in all (HOOYKAAS 1973, pp. 5 and 8), which sometimes 
had a role in state-sponsored rituals (Stuart-Fox 2002, pp. 324 and 326)). 
267 Nagarakrtagama 12.5; PlGEAUD 1960-1963, vol. 4, p. 25. For a map showing 
the location of these quarters within the palace compound (kraton) see Hall 
1996, p. 99. PlGEAUD claims (ibid.) that both are regularly mentioned in the 
preambles of the royal charters of Majapahit. This is so in the Decree Jaya 
Song of c. 1350, the Ferry Charter of 1358, and the undated Charter of Batur 
(PlGEAUD 1960-1963, vol. 1, pp. 104-114 [edition]; vol. 3, pp. 151-164 [transla- 
tion]). They are named in the first after the ministers: the Dharmadhyaksa 
of the Saivas (dharmmadhyaksa ring kasewan), Rajaparakrama, alias Dharma- 
raja, and the Dharmadhyaksa of the Buddhists (dharmmadhyaksa ring kaso- 
gatan) Aryadhiraja Kanakamuni, described as a master of the Buddha's teach- 
ings and grammar (boddhasastrawayakaranaparisamapta). In the second the 
Dharmadhyaksa of the Buddhists has become Nadendra, described in the same 
way (boddhatarkkawyakaranasastraparisamapta) and we learn that the second 
name Dharmaraja of the Dharmadhyaksa of the Saivas is his nama puspapata, 
that is to say, the name he acquired during his initiation through the casting of a 
flower (puspapatah) in accordance with standard Saiva procedure (e.g. Svacchanda- 
tantra 4.62cd: puspapatavasan nama karayet sadhakasya tu 'He should name the 
Sadhaka in accordance with the casting of the flower'; Brhatkalottara f. 91v4 : 
puspapatanusarena samjha Hatpurvato [em. : tatpatrato Cod.] hita 'The [ele- 
ment of the] name before that [such as -siva which indicates the initiate's caste] 
should be [given] in accordance with the casting of the flower'). In the third the 
Dharmadhyaksa of the Buddhists is Aryadhiraja [Kanakamuni], as in the first, de- 
scribed as a master of grammar and the [Buddhist] Tantras (wyakaranatantrapari- 
samapta), and that of the Saivas is Arya Harsaraja, described as a master of logic 
and grammar (nyayawyakaranasastraparisamapta). They are mentioned along 
with a number of other learned men, six in the first, seven in the second, and five in 
the third, referred to as "teachers of Law and settlers of law suits" (dharmmapra- 
wakta wyawaharawicchedaka) in the first and second and as "settlers of law suits 
as valid or not" (nyayanyayawyawaharawicchedaka) in the third. They are no doubt 
the officials referred to elsewhere as the Dharmopapattis (see here p. 105). In the 
first they are (1) Siwanatha, (2) Marmanatha, (3) Smaranatha, (4) Jayasmara, (5) 
Agreswara, and (6) Mumndra. In the second they are (1) Siwanatha, (2) Agreswara, 
(3) Jayasmara, (4) Widyanatha, (5) Siwadhipa, (6) Srighana, and (7) Samatajnana. 
In the third they are (1) Marmanatha, (2) Smaranatha, (3) Mahanatha, (4) a second 
Smaranatha, and (5) Agreswara. Mumndra in the first and Srighana and Samata- 
jnana in the second were Buddhists, a fact already evident from their names but 
confirmed by the charters' reports of their fields of expert knowledge. We learn 
from the first charter that Siwanatha, Smaranatha, and Agresvara were adher- 
ents of the Bhairava sect (bhairawapaksa), that is to say, Sakta Saivas, and that 


Genesis and Development of Tantrism 

to the support of both religions (81.1-2). Moreover, in the opening verse of his 
poem he pays homage to him as Siva-Buddha in human form. 268 

Particularly striking are passages that report the deity-images or temples in 
which the souls of deceased kings had been installed. Ranggah Rajasa (r. 1222- 
1227), was enshrined in two temples, one Saiva and and the other Buddhist, in a 
single temple complex at Kangenengan; 269 and both Saiva and Buddhist priests 
were seated beside king Rajasanagara when he sat in audience after worship- 
ping here. 270 Anusapati (r. 1227-1248) was installed in a Siva image at Kidal; 271 
Visnuvardhana (r. 1248-1268) in a Siva image at Waleri and a Buddha image 

Marmanatha and Jayasmara were adherents of the Saura sect (sorapaksa), that is 
to say, Surya worshippers (see here p. 58). The second and third charters do not 
specify the sects of the judges listed, so that the affiliations of Widyanatha, Siwa- 
dhipa, one of the two Smaranathas, and Mahanatha are unknown. It is striking 
that these judicial boards included no Vaisnavas. The absence of a representative 
of the Rsi sect, often grouped with those of the Saivas and Buddhists as one of the 
three principal denominations in Java (e.g. Arjunawijaya 28.1c: rsi saiwa sogata; 
Kuhjarakarna 22.3c: sang boddhasaiwdrsipaksa), is not surprising. For its fol- 
lowers were forest-dwelling hermits. The Kuhjarakarna associates them with the 
worship of the [Pasupata] pahcakusika; see 23. Id: Iwi glar sogata pahcabuddha rsi 
pahcakusika wiku saiwapahcaka; and Teeuw and Robson 1981, p. 26. See also 
Sanderson 2005a, pp. 374-376. The creation of the post of a Dharmadhyaksa of 
the Buddhists and the inclusion of Buddhists on the judicial board were perhaps re- 
cent developments. For the Sarwadharma charter issued in 1269 during the reign 
of Krtanagara (Pigeaud 1960-1963, vol. 1, pp. 99-103 [edition]; vol. 3, pp. 143- 
150) mentions only a Dharmadhyaksa of the Saivas (Acarya Siwanatha Tanutama: 
mpungku dharmmadhyaksa ri kasewan dang dcdryya siwanatha mapahji tanu- 
tama) together with a board of five other Acaryas, Dharmadewa, Smaradahana, 
Smaradewa, another Siwanatha, and Agraja, not one of whom has an obviously 
Buddhist name (plate 2, recto, 11. 4-7). 

268 Nagarakrtdgama l.lbc: siwa budda sira sakalaniskalatmaka | sang srlparwwata- 
naiha 'The Lord of the Mountain, protector of the unprotected, the holy Siva- 
Buddha, who is both manifest [in physical forml and transcendent'. The Lord of 
the Mountain (sriparwwatandtha) addressed in this verse has been understood, im- 
plausibly, as Siva. I am entirely persuaded by the evidence presented by Supomo 
(1972; 1977, pp. 69-82) that it is the king that is intended in this and the opening 
verse of Mpu Tantular's Arjunawijaya, where the Lord of the Mountain, in this case 
called Parwwatarajadewa, is identified as the physical manifestation of the ultimate 
reality that is the Buddha (1.1b: sang saksat paramdrthabuddha). 

269 Nagarakrtdgama 40. 5d: sang dindrmmadwaya ri kagnangan ssewaboddeng usdna. 
Pigeaud translates dindrmmadwaya as 'a double dharma (religious domain)' 
(1960-1963, vol. 3, p. 46) and ROBSON (1995, p. 5) as 'a double temple'. I do not 
see that the expression, which is equivalent to Skt. dharmadvayam, conveys any- 
thing more than the fact that there were two temples. Cf. Santoso 1975, p. 54. 

270 Nagarakrtdgama 36.2b: para wiku sai sogata dryya ndligih iniring nirekhi tan 

211 Nagarakrtdgama 41. Id: pradipa *siwabimba (Kern : simbha Pigeaud) sobhita 
rikang sudharmma ri kidal. 


The Saiva Age 

at Jajaghu; 272 Krtanagara, r. 1268-1292, who is depicted as a devout initiated 
Tantric Buddhist and described after his death as "liberated in the world of Siva- 
Buddha", 273 and was installed in a Siva-Buddha in "his own place" and, with 
his queen, Vajradevi, in a Buddhist image combining Vairocana and Locana at 
Sagala. 274 Krtarajasa Jayavardhana (r. 1293-1309) was installed in a Buddha 
in the palace and a Siva at Simping, 275 and Jayanagara (r. 1309-1326), who is 
described as having returned to the world of Visnu, 276 in Visnus in the royal com- 
pound, Shila Petak, and Bubat, and in a Buddha in the form of Amoghasiddhi 
in Sukhalila. 277 We also learn that there was a temple founded by Krtanagara 
at Jajawa, located at the foot of the sacred mountain Kukuwus, which was Saiva 
but had a Buddhist pinnacle and contained a Siva with an image of Aksobhya 
above its crown, and that both Buddhists and Saivas worshipped in it. 278 The in- 

272 Ndgarakrtdgama 41.4b: dinarmma ta sire waleri siwawimbha len sugatawimbha 
mungwing jajaghu. 

273 Ndgarakrtdgama 43.5c: sang mokteng siwabuddaloka . His commitment to Bud- 
dhism is indicated in 42.3c (samaya len brata mapageh apdksa sogata) and 
43.2a (bhakti ri pada sri sakyasimhdsthiti). As for his involvement in Tantric 
Buddhism we learn that he received Buddha consecration (jindbhisekah) and 
was then given the name Jnanavajresvara (43.2bc: lumrd ndma jindbhisekanira 
sang sri jndnabajreswara), that he devoted himself to Tantric worship fol- 
lowing the otherwise unknown Subhutitantra as his principal guide (43.3b: 
mukyang tantra subhuti rakwa tinngbt hempen), and that he celebrated the 
esoteric Tantric ritual known as ganacakram (43.3d), an indication that his 
Tantrism was that of the Guhyasamdja or one of the Yogimtantras. His 
initiation-name appears in the forms Jnanasivavajra and Vajrajnanasiva in the 
Sanskrit inscription (Kern 1910) on the lotus-cushion of an image of him- 
self in the form of the Mantranaya deity Mahaksobhya installed at Simpang 
in Surabaya in 1289 (vv. 12—13: srijndnasivavajrdkhyas cittaratnavibhusanah | 
jndnarasmivisuddhdhgas sambodhijhdnapdragah || subhaktyd tarn pratisthdpya 
svayam purvam pratisthitam \ smasdne vurarendmni mahdksobhydnurupatah; 19d: 
vajrajnanasiva + +). All three forms of the name have the appearance of a Saiva- 
Buddhist hybrid. 

274 Ndgarakrtdgama 43. 5d: rihke sthdnanirdn dinarmma siwabudddrcca halp no- 
ttama; 43.6: hyang werocana locana Iwiriran ekdrcca prakdseng prajd. 

275 Ndgarakrtdgama 47.3b-d: drdk pinratista jinawimbha sireng puri jro | antahpura 
ywa panlah rikanang sudarmma saiwdpratista sira teko muwah ri simping. 

276 Ndgarakrtdgama 48.3a: sang nrpati mantuk ing haripada. 

277 Ndgarakrtdgama 48.3bcd: sighra sirdn dinarmma ri dalm purdrccanira 
wisnuwimbha parama | len ri sild ptak mwang i bubat pada pratima wisnumurtty 
anupama ring sukhalila tang sugatawimbha sobhitan amoghasiddhi sakala. His 
installation in Visnus is without parallel among the Singhasari-Majapahit kings; 
see Pigeaud 1960-1963, vol. 4, p. 141. However, the kings of Kadiri, the principal 
court of East Java through the tweflth century to c. 1222, were devotees of this god. 
Most were described as his embodiments (de Casparis and Mabbett 1992, p. 327) 
and his incarnations are central to the literary epics (kakawin) of the Kadiri court 
(Hall 2005, pp. 2 and 8). 

278 Ndgarakrtdgama 56.1b— 2c: kirtti sri krtanagara prabhu yuyut nareswara sira | 


Genesis and Development of Tantrism 

timate co-existence of the two traditions is also apparent in the intertextuality of 
religious texts in Java, as has been demonstrated for the Saiva Jnanasiddhanta 
and the Tantric Buddhist Sang hyang Kamahdydnikan and Kalpabuddha. 279 It 
is also seen in the great frequency with which the Mahayana-Buddhist concept of 
emptiness (sunyatd) is incorporated in Javanese Saiva sources through the inclu- 
sion of the terms sunya and sunyatd among those used to characterize the high- 
est reality 280 in the readiness of the redactors of Saiva liturgies to supplement 
sets of Saiva elements with Buddhist elements when they needed to make up a 
total for the sake of the numerical correspondence, 281 and in the fact that the 
Kunjarakarna of the Buddhist Mpu Dusun the supreme Buddhist deity Vairo- 

tekwan rakwa sirangadistita sarlra tan hana waneh etunyang dwaya saiwa bodda 
sang amuja nguni satatd || chinang candi ri sor kasaiwan apucak kaboddan i 
ruhur mwang rijro siwawimbha sobhita halpniraparimita | aksobhyapratime ruhur 
mmakuta tan hanolyantika 'It was a temple (kirtih) of Lord Krtanagara, the king's 
great-grandfather. He himself established it. Hence both Saivas and Buddhists 
have from the beginning always conducted the worship. The sign is that the temple 
is Saiva in its lower section and Buddhist above. Inside is a beautiful Siva image 
and above an image of Aksobhya as (on?) its crown. Of there is no doubt'. On the 
significance of the Saiva-Buddhist fusion seen in Krtanagara in both inscriptions 
and literary works see Hunter 2007. 

279 See Soebadio 1971, pp. 12-19 and 55-57 for evidence of this intertextuality; also 
for a general treatment of the co-existence of the two traditions in Java ZOETMUL- 
der in Stohr and Zoetmulder 1968, pp. 262-314. 

280 See, e.g., Jnanasiddhanta 3.2-3: nadas ca liyate sunye sunyam eva tu jayate 
| sunydc chunyataram vapi atyantasunyalaksanam || sthulam sakalatattvam ca 
suksmam sakalaniskalam | param niskalasunyam ca urdhvatyurdhatisunyakam; 
8.3: sthulam sabdamayam proktam suksmam cittamayam bhavet \ param 
cittavirahitam cittam tyaktvatisunyata; Ganapatitattva 2: svaso nihsvasah 
samyoga atmatrayam iti smrtam | trisivam tripurusatvam aikatmya eva sunyatd; 
23: hrdayastham sadasivam hrdayante guhyalayam I sunyatisunyam cinty- 
ate param kaivalyam ucyate; Mahajnana 62: suryakotisahasramsu hrdayam 
vimalam subham \ hrdayante padam sunyam param kaivalyam ucyate; 83: ratris 
ca prakrtir jheya ravis ca purusas tathd \ dyutis ca va mahadevah sunyam 
ca paramah sivah. I consider it highly probable that these Sanskrit works 
are Javanese creations. Some of the verses can be found in Indian Saiva 
sources: Wrhaspatitattwa 53 and Ganapatitattwa 3 < Rauravasutrasamgraha 7.5; 
Jnanasiddhanta 19.5 and and Ganapatitattwa 43 < Kirana 1.23; Wrhaspatitattwa 
7-10 < Svayambhuvasutrasamgraha 4.3-6. But these are surprisingly few, and 
the works contain several doctrinal elements that are alien to known Indian tra- 
ditions. Moreover, the deviations from strict Sanskrit usage found in them seem 
to me not to be characteristic of the registers of the language seen in Indian Saiva 
scriptural texts. The same is true of the frequent deviations from the correct form 
of the Anustubh in the second and fourth Padas: e.g. Ganapatitattwa Id, 16d, 48d, 
49b, 49d, 54b, 54d, 55b, 59b, 59d; Mahajnana lid, 37b, 38d, 42b, 61b, 73b, 74b, 
78b, 78d, 80d; Wrhaspatitattwa 3b, 6b, 6d, 12b, 20d, 23b, 24b, 25b, 63b, 72d. This is 
extremely rare in Indian Saiva texts. 

281 See the example of this cited in Sanderson 2005a, p. 377. 


The Saiva Age 

cana is made to equate the divine pentads of the Saiva and [Pasupata] Rsi sects 
with the five Tathagatas, teaching this in the context of an assertion that he is 
the ultimate reality that assumes the form both of the Buddha and of Siva, 282 
and that it is because the followers of the three sects fail to understand this un- 
differentiated ground that they dispute with each other for the pre-eminence of 
their respective Gods. 283 The same idea is seen in the works of the Buddhist Mpu 
Tantular. In his Arjunawijaya he has the priest of a Buddhist temple-complex 
(boddhadharma) explain to Arjuna that its central diety Vairocana is one with 
Sadasiva, that its four ancillary deities, the directional Tathagatas Aksobhya, 
Ratnasambhava, Amitabha, and Amoghasiddhi, are one with Rudra, Brahma, 
Mahadeva, and Visnu respectively, 284 that there is no distinction between the 
Buddha and Siva, 285 and that therefore it is the king's duty to support all three 
sects, the Buddhists, the Saivas, and the Rsis. 286 Later, in his Sutasoma, Mpu 
Tantular states that the Buddha and Siva are "different but one" (bhinneka tu- 
nggal ika), the famous formula that has been adopted as its official national 
motto by the modern state of Indonesia, as two manifestations of the ultimate 
reality of the former. 287 

282 Kunjarakarna 23. Id: livir glar sogata pancabuddha rsi pancakusika wiku saiwa 
pancaka 'As the Buddhists have the five Buddhas, the Rsis have the pentad of 
Kusika and the Saivas a pentad of their own'; 23.4bcd: ngwang wairocana buddha- 
murti siwamiirti pinakaguru ning jagat kabeh \ ndham donkw ingaran bhatara 
guru kaprakasita teka ring sardt kabeh | anghing bydpaka ring samastabhuwandku 
juga warawisesadevatd 'I, Vairocana, am embodied both as the Buddha and as Siva, 
and am accepted as Guru by all. Therefore it is I that am Bhatara Guru, famed 
among all men, and it is I, as the highest deity, that pervade all the worlds.' 

283 Kunjarakarna 22.3. 

284 Arjunawijaya 26.4-27.1 

285 Arjunawijaya 27.2abc: ndah kantenanya haji tan hana bheda sang hyang | hyang 
buddha rakwa kalawan siwa rajadewa | kalih sameka sira sang pinakestidharma. 

286 Arjunawijaya 30.1-2. 

287 Sutasoma 139.5: hyang buddha tan pahi lawan siwarajadewa | rwanekadhatu 
winuwus warabuddhawiswa | bhinneki rakwa ring apa n kena parwanosen 
mangka ng jinatwa kalawan siwatattwa tunggal \ bhinneka tunggal ika tan hana 
dharma mangrwa. This has been translated by Supomo (1977, p. 7) as follows: 
"The god Buddha is not different from Siwa, the lord of the gods. The excellent 
Buddha, the all-pervading, is said to be two different dhdtu. Yet although these two 
dhatu are different, how is it possible to differentiate between them at a glance? In 
the same manner, the reality that is Jina and the reality that is Siva are one; they 
are different yet they are one, for there is no duality in the dharma". Comment- 
ing on "the two different dhatu" mentioned in this verse (fn. 9) Supomo take them 
to be the two Mandalas, the Garbhadhatu and the Vajradhatu of the Mahavairo- 
canabhisambodhi and Sarvatathagatatattvasamgraha respectively. This reading 
is an error in my view. It does not accord with context, which requires that the 
two be the realities of the Buddha (jinatwa) and Siva (siwatattwa) respectively. As 
I understand it, the passage is saying that the Lord Buddha is both the Buddha 


Genesis and Development of Tantrism 

The Development of Tantric Buddhism Through the Adoption and 
Adaptation of Saiva and Sakta Saiva Models 

The Parallel Repertoire of Rituals 

Now, this co-existence of Buddhism and Saivism under royal patronage was 
surely facilitated by the fact that the form of Buddhism adopted and developed 
was one that had equipped itself not only with a pantheon of ordered sets of 
deities that permitted such subsumptive equations but also with a repertoire 
of Tantric ceremonies that parallelled that of the Saivas and indeed had mod- 
elled itself upon it, offering initiation by introduction before a Mandala in which 
the central deity of the cult and its retinue of divine emanations have been in- 
stalled, and a system of regular worship animated by the principle of identifi- 
cation with the deity of initiation (devatahamkarah, devatagarvah) through the 
use of Mantras, Mudras, visualization, and fire-sacrifice (homah); and this was 
presented not only as a new and more powerful means of attaining Buddha-hood 
but also, as in the Saiva case, as enabling the production of supernatural ef- 
fects (siddhih) such as the averting of danger (santih), the harming of enemies 
(abhicarah), and the control of the rain (varsapanam and ativrstidharanam), 
through symbolically appropriate inflections of the constituents of these proce- 
dures. The latter is particularly important from the point of view of Buddhism's 
relations with its royal patrons, since such rituals enabled it to match the Saivas 
by promising kings more tangible benefits than the mere accumulation of merit 
through the support of the Buddha, his teaching, and the Sahgha. We have seen 
an example of such ritual for the protection of the state in Taranatha's report 
of the programme of Tantric fire-sacrifices performed at Vikramasila under the 
direction of Buddhajnana during the reign of Dharmapala (r. c. 775-812) to en- 
sure the longevity of the Pala dynasty; 288 we have another example in the case of 
Klrtipandita, a Mahayana-Buddhist scholar and Tantric expert who according to 
the Vat Sithor stele inscription became the Guru of the Khmer king Jayavarman 
V (r. 968-1001) and was engaged by him to perform frequent fire-sacrifices in 
the palace for the protection of the kingdom; 289 and the Javanese Prapanca tells 
us that the purpose of king Krtanagara's adherence to Tantric Buddhism was 

and Siva, whereas Supomo's reading makes Mpu Tantular espouse a doctrine of 
absolute equality between the two religions within a reality beyond both. This is 
intrinsically implausible in a Buddhist work. My reading makes his view exactly 
that expressed by Mpu Dusun in 23.4bcd of the Kunjarakarna cited and translated 
above: '"I, Vairocana, am embodied both as the Buddha and as Siva". 

288 See here p.93. 

289 K. Ill, Ccedes 1937-1966, vol. 6, pp. 195-211, v. 36. See Sanderson 2005a, 
pp. 427-428. 


The Saiva Age 

to increase his people's prosperity and the stability of his realm, and that its 
reward was the undiminished and undivided sovereignty (ekachattra) of his de- 
scendants. 290 

The adoption of the Saiva practice of Mandala initiation created a further 
line of access to patronage and was propagated vigorously, as it was by the 
Saivas, as a means of the recruiting of social elites both in the subcontinent and 
beyond. 291 Among the Buddhist Tantras at least two major texts teach rituals of 
initiation, or consecration (abhisekah) as it is called in these sources, in which 
it is kings in particular and royalty in general that are envisaged as the pri- 
mary initiands. These are the Manjusriyamulakalpa and the Sarvadurgatipari- 
sodhanatantra. 292 In the former this is so for the principal Kalpa of the text. In 
the latter it is characteristic of initiation into the secondary Mandalas of the four 
Great Kings and the ten Guardians of the Directions taught in the Uttarakalpa. 
The sections dealing with these Mandalas specify the king as the principal con- 
secrand, teach little or no required subsequent practice, and promise benefits 
that apply principally to him, namely the protection of himself and his kingdom 
and the destruction of the kingdoms of his enemies. The monarch is not men- 
tioned in the treatments of initiation given in the Mahdvairocandbhisambodhi 
and Sarvatathdgatatattvasamgraha, the two great Tantras that were translated 
into Chinese in the early eighth century to form the basis of the Way of Mantras 
there and in the Japanese Shingon and Tendai sects. But the ninth-century In- 
dian authority Anandagarbha brings this aspect of the religion to the fore in his 
Sarvavajrodaya, an influential manual that sets out detailed practical guidance 
for the performance of the initiation ritual taught in the second of those texts but 
draws heavily on the more detailed treatment in the first. For when he teaches 
the preparation of the Mandala he prescribes a range of sizes beginning with that 
appropriate for the initiation of the monarch. In his case each of the sides should 
measure one hundred or fifty cubits (about 40 and 20 metres), in the case of a 
feudatory (sdmantah) or major feudatory (mahdsdmantah) fifty or twenty-five, 
in the case of a wealthy merchant (sresthi) or international trader (sdrthavdhah) 
twenty-five or half of that, and in the case of an ordinary practitioner (sddhakah) 

290 Nagarakrtdgama 42.3d: tumlrwa sang atitardja ring usdna magehakna wrddining 
jagat; 43.3c: pujd yoga samddi pinrihiran amrih sthityaning rat kabeh; 43.4cd: 
darmmestdpageh ing jinabrata mahotsdheng prayogakriya nahan hetuni tusni tus- 
nira padaikaccatra dewaprabhu. 

291 On the adoption by the Buddhists of the practice of royal initiation and its propaga- 
tion in India, Tibet, Mongolia, China, Japan, and Southeast Asia see Sanderson 
forthcoming a. 

292 Manjusriyamulakalpa, p. 32, 11. 21, 23, and 28-30; Sarvadurgatiparisodhana- 
tantra, sections 47b, 48a, and 49a. 


Genesis and Development of Tantrism 

twelve or six (about 5 or 2.5 metres). 293 

The Mantranaya also followed the example of the Saivas by devising Tantric 
ceremonies for patrons in the public domain: for the consecration (pratisthd) of 
temple images (pratimd), paintings of deities on cloth (patah), manuscripts of 
sacred texts (pustakam), monasteries (viharah), shrines (gandhakutl), Caityas, 
reservoirs ipuskarinyadi), gardens and the like (drdmddi). It also adapted the 
Saiva procedures for funerary initiation to produce a Tantric Buddhist funeral 

293 Sarvavajrodaya , f. 29r5-29vl: evam krtvd purvasevdm mandalam dlikhet. . . . rdjho 
hastasatam pahcdsaddhastam vd sdmantamahdsdmantdndm pahcdsat paiica- 
vimsatihastam vd sresthinah sdrthavdhasya vd pancavimsatim tadardham vd 
sddhakdndm dvddasahastam saddhastam vd. 

294 The details of this wide repertoire of the rituals that a Tantric Buddhist offi- 
ciant (Vajracarya) was called on to perform are set out in a number of man- 
uals that are closely comparable to the Paddhatis of the Saivas, notably the 
Kriydsamgrahapanjikd of Kuladatta (Tanemura 2004b), the Vajrdvali of the 
great Abhayakaragupta of Vikramasila (1064-1125 according to the chronologi- 
cal tables of Sum pa mkhan po Yes shes dpal 'byor [1704-1788]; works dated 
in the twenty-fifth, thirtieth, and thirty-seventh years of the reign of Ramapala 
[c. 1072-1126]; Vajrdvali written before the first of these; see BtJHNEMANN and 
TACHIKAWA 1991, pp. xiv-xvi), which adds procedures for the consecration of reser- 
voirs, gardens, and the like (A, f. 2rl in the list of topics: pratimddipratisthd 

| puskarinyddipratisthd | drdmddipratisthd), and the Acdryakriydsamuccaya of 
Mahamandalacarya Jagaddarpana, which incorporates much of the Vajrdvali but 
adds some new material, notably a final section on the funeral ritual for a de- 
ceased Vajracarya (nirvrtavajrdcdrydntyestilaksanavidhih; B, ff. 240v7— 244v4), 
which is an unacknowledged incorporation of the whole of the Mrtasugatiniyojana 
of Pandita Sunyasamadhivajra (less its two colophonic verses). One other text giv- 
ing a Tantric funeral procedure survives in Sanskrit, the Antasthitikarmoddesa, 
at the end (ff. 15r8-15vll) of the Guhyasamdja -based Mandalopdyikd of 
Mandalacarya Padmasrimitra of the Khasarpana monastery (f. 15vl0-ll: samdptd 
ca mandalopdyikd | krtir iyam khasarpaniyamandaldcdryapadmasrimitrasya). On 
these texts and the incorporation of the Mrtasugatiniyojana by Jagaddarpana see 
Tanemura 2004a and 2007. On the Saiva prototype of funerary initiation see 
Sanderson 1995a, pp. 31-33 and, for its adaptation, the Mrtoddhdradiksd, in 
which a simulacrum is substituted for the body of the deceased, 2005b, pp. 264-267. 
A fourteenth-century Paddhati for this Mrtoddhdradiksd survives in ff. 88vl-91rl 
of the Gurupustikd of the Kashmirian Rajanaka Sitikantha. In an earlier publi- 
cation (Sanderson 2007a, p. 395, fn. 549) I proposed that this work, then known 
to me only indirectly from the Rdjdnakavamsaprasamsd of his patrilineal descen- 
dant Rajanaka Ananda, who reports that it was composed at the request of [king] 
Samgramasimha, might be preserved in a Sarada manuscript listed with this ti- 
tle as belonging to the Sayaji Rao Gaekwad Central Library of the Banaras Hindu 
University (MS CN. 4115). I can now report that this is indeed a manuscript of 
that work and, as far as I am aware, its codex unicus. The name of the author is 
confirmed on f. lvll-12: karmdnupurvismrtaye kesdmcid upayoginim | sitikanthas 
samasyaindm vidhatte gurupustikdm; and the claim that he wrote at the re- 
quest of Samgramasimha is confirmed on f. 13vl5-14rl: asmdkam kulasisyena 
srisangrdmamahibhujd | abhyarthitdndm diksdrtham ayam paddhatidohadah. I 
am very grateful in this matter to my former pupil Christopher Wallis, who after 


The Saiva Age 

rite (antyestih) 294 for initiates, 295 in which, as in the Saiva case (antyestidiksa), 
the officiant draws the consciousness (jnanam) of the deceased back into the 
corpse from the other world, takes it again through the initiatory process of con- 
secration and the rest (abhisekadi) before a Mandala, 296 and then sends it out 
through the top of the head to ascend to liberation or a pure Buddha-field such 
as Sukhavati. 297 

reading my remark that I had not yet seen the manuscript very kindly acquired and 
sent me scans of it. 

295 According to Padmasrimitra the ritual is to be done for Acaryas and others who have 
practised the meditation-rite of Vajrasattva or some other Tantric deity; f. 15r8, 
v. 1: mrtdcdryddisattvd ye vajrasattvddiyoginah | vaks<y>e cdntasthite<h> krtyam 
tesdm mdrganidarsandt . It may be done for a man or a woman; f. 15rl0— 11, v. 
9ab: purusatanu<m> nirupydtha striyo vd samyag eva hi. Sunyasamadhivajra does 
not specifiy those for whom it is intended. But Jagaddarpana adds a preamble to 
Sunyasamadhivajra's text in which he restricts it to Vajracaryas; f. 240v7: adhund 
parinirvrtavajr deary asarlrasydntestividhir ucyate. 

296 Mandalopdyika , f. 15rl4, vv. 21c-22b: tato vijnanam dnlya mantramudrd- 
nuyogatah || ahkusyddyaih pravesydtha dadydt sekddikam punah 'Then having 
drawn down the consciousness [of the deceasedl by means of the Mantras and 
Mudras, and having caused it to enter [the corpse] by means of the Mudras begin- 
ning with the Hook, he should again give it the consecrations and the rest'; Mrta- 
sugatiniyojana, f. 2r3-4: tato nayet suraktavarnam (conj. [Tib. mdog dmar gsal 
ba] : suraktamsvadhdm) paralokasamsthitam jnanam dharmamukhdkrti yad vd 
nivdtaniskampadipanibham \ dnitam taj jnanam mrtasya hrdaye pravesayet sirasd 
'Then he should draw down the consciousness [of the deceased] that is in the world 
beyond, [visualizing it as] bright red in colour or with the shape of the letter A (the 
dharmamukham), resembling the unflickering flame of a lamp in a windless place. 
When that consciouness is nigh he should cause it to enter the heart of the deceased 
through [the top of] his head'. According to the Mandalopdyika, the Acarya should 
trace and worship the Mandala, offer a Bali, and then place the corpse at its east 
gate with its head to the south; f. 15rl2-13, vv. 12-13b: same visuddhabhubhdge 
gomayenopalepite \ mandalam catusram vai kdrayet tatra samkiret || suklam pitam 
rajo vdpi tatra padma*daldstakam (conj. : daldbhakam Cod.); f. 15rl3, vv. 18c— 19: 
uttardbhimukho mantri sampujya mandalam balim || dattvdrghddikam caiva sam- 
sddhya mandalam krti | sthdpayen mandaladvdri prdcydm tu daksindmukham . 

297 In the Mandalopdyika 's prescription the Acarya visualizes that the purified con- 
sciousness of the deceased is drawn out of the corpse by a multitude of rejoic- 
ing deities filling the sky and placed by them in a world such as Sukhavati in- 
habited by Buddhas and Bodhisattvas; f. 15v2-3: 28 sambuddhabodhisattvddi- 
virinwlravrndakaih | siddhagandharvabhujagaih surair vidyddharair api || 29 
purnam nabhastalam viksya *nipatatpuspavrstikam (nipatat em. : nipatatah 
Cod.) | tad divyadundubhidhvdnamuraja*mardaladhvani (mardala conj. : mu- 
rdata Cod.) || 30 ucchatavenuvinddimadhurasvdrabhusanam \ taddnandasuvistdrdt 
kurvadbhir nrtyam ujjvalam \\ 31 tair dkrsya ca vijnanam sukhdvatyddikdhvaye 
| sthdpitam lokadhdtau hi buddhabuddhdtmajdsraye. The procedure of the 
Mrtasugatiniyojana differs here; f. 3rl-3: tad anu *kusdgre (em. [Tib. ku sha'i 
rise mo la] : kusdgram Cod.) *mantrl (em. [Tib. sngags pas] : mantrai Cod.) 
vibhdvya tiksnaika*sucikam vajram (corr. : sucikavajram Cod.) | niksipya va- 
jrarandhre dhydydt tad dahanasamkdsam || tad anu samdhitacitta<s> taddhrdi 
vinyastavisphuraj jnanam \ samcodayej *jvaladbhir vajrdgrair mdrutoddhutaih 


Genesis and Development of Tantrism 
The Mahdvairocanabhisambodhi, the Manjusriyamulakalpa, and Buddhaguhya 

That this transformation of the Mahayana had been achieved by absorbing 
and adapting non-Buddhist practices was evident from the beginning. For the 
Mahdvairocanabhisambodhi, our first major Buddhist Tantra, 298 later classified 
as the principal work of the Caryatantra class, was conscious that it would be 
accused of just this: 

O [Vajrapani,] Lord of the Yaksas, in time to come there will arise people of in- 
ferior understanding and no faith who will not believe this teaching. They will 
dissent and have many doubts. They will hear it but they will not take it to heart 
and they will refuse to put it into practice. Being themselves unworthy they will 
bring others too to ruin. [For] they will say that this is not the teaching of the 
Buddhas but belongs to the outsiders. 299 

(em. [Tib. rdo rje rise nas Hung gis bskyod pa yi 'bar ba mams kyis]: jvalad- 
bhivajragraumarutoddhrtair Cod.) || *udgacchad tad (corr. [Tib. de ni 'phar 
bar] : udgacchantam Cod.) dhyayad dahanarci<h>sprsyamanaparadavad | 
*urdhvagnena (?) (Tib. steng gi sgo nas) vimuktim buddhaksetram visuddham 
va 'Then the Mantrin should take a blade of Kusa grass, visualize a sharp one- 
pointed Vajra at its tip, place [that tip] at the aperture of the [corpse's] penis and 
imagine that it is burning. Then concentrating his mind he should cause the shin- 
ing consciousness that he has installed in the heart [of the corpse] to be driven 
[up from the heart] by blazing wind-fanned Vajra-points and he should visualize 
it rising to liberation or a pure Buddha-field through the upper [aperture], like [a 
drop of] quick silver touched by tongues of fire'. The 'upper' is one of nine aper- 
tures through which consciousness can leave the body at death (utkrantih). It is 
located at the top of the head and is called 'the golden door' (kanakadvaram) by 
Bhavabhatta in his commentary on the Catuspithatantra (Catuspithanibandha), f. 
52r2: urdhveti kanakadvarena yada gacchati tada maranad urdhvam sighram eva 
gater gatyantaram visistam gacchati. The point of exit depends upon the destiny of 
the deceased. This is the best. According to Sunyasamadhivajra consciousness that 
exits at death through this aperture goes to the Immaterial World (drupyadhdtuh): 
sirasarupyam gacchet (f. 3r4). This idea that consciousness may leave the body 
through various exits in accordance with its destiny is found widely in Brahmani- 
cal sources. Early Buddhist sources speak rather of consciousness ceasing at death 
at these points in the body; see Abhidharmakosabhasya on 3.43abc. Vasubandhu 
says there that in the case of Arhats their consciousness disappears in the heart 
according to some and in the head according to others: arhantah \ tesdm api hrdaye 
vijnanam nirudhyate \ murdhnity apare. 

298 See here p. 101. 

299 rNam par snang mdzad chen po mngon par byang chub pa'i rgyud, f. 177rl-3: de la 
gsang ba'i bdag po ma 'ongs pa'i dus na sems can bio zhan pa ma dad pa gang dag 
bstan pa'i de la dad par mi 'gyur zhing yid gnyis dang som nyi mang ba | thos pa 
tsam snying po ma 'dzin pa | sgrub ma la mi phyogs pa dag 'byung bar 'gyur te | de 
dag ni bdag nyid kyang ma rung la gzhan yang phung bar byed pa yin no | 'di shad 
du 'di ni phyi rol pa mams la yod de \ sangs rgyas mams kyi gsungs pa ni ma yin 
no zhes smra bar 'gyur gyi. 


The Saiva Age 

The Mahjusriyamulakalpa, another early Buddhist Tantric text, 300 assigned 
to the lowest class of Mantranaya texts, known as the Kriyatantras, is more 
explicit in this regard; and it has good reason to be so since it contains in its 
chapters 47-49 an assimilated version of the cult of Tumburu and his four sis- 
ters, that is to say, the cult of the vamasrotah division of the Saiva Vidyapitha, 
describing the Mantras of these deities as the highest and most secret of all 
the non-Buddhist (laukika-) Mantras. 301 Moreover, it teaches that any of the 

300 The date of this text is obscure. Matsunaga (1985) is of the opinion that the 
9th chapter, on applications of the Ekaksaramantra, was in existence before the 
Chinese translations T 1181 of a.d. 702 and T 1182 of a.d. 703. He also in- 
forms us (ibid. ) that the first ninety percent of the Chinese translation of the 
Garudapatalaparivarta (T. 1276), produced at some time between 746 and 774, is 
identical with the first sixty percent of the 41st chapter of the Mahjusriyamulakalpa 
as edited. The translation is attributed to Amoghavajra (705-774), but Mat- 
sunaga observes (ibid. ) that only the first part of the common text is in keeping 
with his other translations, the latter part containing elements such as human hair, 
beef, and skull-cups, which taken together are altogether alien to his Mantranaya. 
He strengthens the hypothesis that only the first part of this translation is by 
Amoghavajra with the evidence of the Go-shorai mokuroku, a catalogue of the Bud- 
dhist texts brought from China to Japan by Kukai in 806, which lists this text as oc- 
cupying three sheets, a third of the length of T 1276. The prophetic history of Indian 
Buddhism, the Rajavyakarana, chapter 53 of the published Majusriyamulakalpa, 
cannot be earlier that the late eighth century since it knows of the Pala king Gopala 
(r. c. 750-775) (53.628; and 53.816: tatah parena *bhupalo gopalo [em. : bhupala 
gopala Ed.] dasajivinah | bhavisyati). Since it does not mention his successor 
Dharmapala it is unlikely to be later. 

301 Manjusriyamulakalpa, introductory prose before 47.1: sarvalaukikamantranam 
sdrabhutatamam paramarahasyam. The position within Saivism assigned by this 
text to the cult of the four sisters suggests that, though later largely eclipsed by 
other traditions of the Vidyapitha, it was once pre-eminent; and this is also cir- 
cumstantial evidence in favour of the hypothesis proposed above (p. 50) that this 
cult was one of the earliest, perhaps the earliest, of the esoteric Saiva systems. 
There is certainly much other evidence of its early centrality. As we have seen, 
it was known to Dharmakirti (here p. 50), and a 6th-century manuscript of one 
of its texts survives amid the otherwise Buddhist Gilgit manuscripts (here p. 50). 
The Visnudharmottara shows knowledge of only two Saiva deity-systems in its 
section on iconography: the Saiddhantika and this (3, Adhyaya 66, teaches the 
iconography of Tumburu and his sisters). The Advaitin Sankara in his Gitabhasya 
on Bhagavadgita 9.25, in which it is said that those who worship the Spirits 
(bhutejyah) reach the Spirits (bhutani yanti) [when they die], glosses bhutani as 
vinayakamatrganacaturbhaginyadlni 'such as Vinayaka, the Mothers, and the Four 
Sisters'. On his date, probably eighth century, see Harimoto 2006. These deities 
were also incorporated in the traditions of Mandalas of the Nayasutra and the 
Mahavairocanabhisambodhi that reached the Far East in the eighth century (see 
Sanderson 2001, p. 8, fn. 5). Their cult was the basis of the Saiva ritual performed 
to inaugurate the kingdom of Angkor at the beginning of the ninth century (ibid. 
and 2005a, pp. 355-358); and there too, where the Mantramarga was preserved in 
an early form, we see only the Siddhanta of its earliest texts and this cult. This 
co-existence is also evident in the Saiva liturgies of Java and Bali, which are of 
Saiddhantika character but incorporate these deities (see Goudriaan 1973 and 


Genesis and Development of Tantrism 

Mantra-procedures taught in the Saiva and Garuda Tantras 302 will be effec- 
tive if applied by Buddhists in the Mandala of these converted deities. 303 Thus 
the Buddhists envisaged by this text have the whole array of Saiva Mantras 
at their disposal; and this position, so surprising from the conventional Bud- 
dhist standpoint, is justified by the claim that what people have come to refer to 
as the Saiva, Garuda, and indeed Vaisnava Tantras are in fact Buddhist, since 
they were first taught by Mahjusri in this "vast Kalpa", that is to say, in the 
Manjusriyamulakalpa or, more probably, in a hypothetical proto-text of which 
the actual text was thought to be an abbreviated redaction: 304 

I have taught this Mantra [of Siva] which together with the trident Mudra 
destroys all demons, out of my desire to benefit living beings. Those living on the 
earth will say that its ancient Kalpa, that I taught in former times, was taught 
by Siva. [Butl the various excellent extensive [Kalpasl in the Saiva Tantras are 
in fact my teachings. 

The extensive Kalpas that have been related in the Vaisnavas Tantras were 
taught by Manjughosa for living beings who could only be trained by [this] 
device. 306 

All the extensive Kalpas taught in the Garuda Tantras were taught by me in 
order to benefit living beings. 306 

It was I that first taught, in this vast Kalpa, everything that the inhabitants 
of earth without exception refer to as the teaching of Siva. It was only later 
that others taught in the various texts [considered to be taught by him] the 
Kalpamantras of the wise Siva Tumburu the Trader. 307 

Sanderson 2005a, p. 373-374, fn. 76). 

302 On the Saiva Garudatantras see here p. 46 and Slouber 2007. 

303 Manjusriyamulakalpa 47.98c— 99b, 102ab, 103ab: yavanti saivatantre 'smim 
ye tantre capi garude || brahmadyair rsimukhyais ca ...pujita kalpavistara 
visnurudrasavasavaih | . . . tasmin mandale *yojya (conj. : yojya Ed.) siddhyantlha 
na samsayah 'All the extensive Kalpas that have been taught in this Saivatantra 
and, moreover, in the Garuda, and worshipped by Brahma and others, by the lead- 
ing Rsis, ... by Visnu, Rudra, and Indra, will be mastered if applied in this Mandala. 
Of this there is no doubt. 

304 Manjusriyamulakalpa 2.32— 34b: esa mantro maya proktah sattvanam hitakamyaya 
| sulamudrasamayuktah sarvabhutavinasakah || 33 yan maya kathitam purvam 
kalpam asya puratanam | saivam iti vaksyante sattva bhutalavasinah || 34 vividha 
gunavistarah saivatantre mayoditah. 

305 2.31c-32b: ya eva vaisnave tantre kathitah kalpavistarah || upayavaineyasattvanam 
manjughosena bhasitah. 

306 2.37: yavantah garude tantre kathitah kalpavistarah \ te mayaivoditah sarve 
sattvanam hitakaranat . 

307 47.53—54: sarvam saivam iti khyatam sarvair bhutalavasibhih \ mayaiva nigaditam 


The Saiva Age 

If this is so, then the text has disarmed criticism that the Mantra-procedures 
that are presented as properly Buddhist in this text bear a suspiciously close re- 
semblance to the non-Buddhist in their liturgical morphology. For if the Omni- 
scient has revealed all forms of religion in consideration of the differing mental 
dispositions of his manifold audiences, then there is no reason at all why he 
should not in his wisdom have taught Tantric practice for Buddhists as well as 
for outsiders. The strict division between the Buddhist and the non-Buddhist 
has dissolved within a higher Buddhist intertextual unity. Indeed this very ar- 
gument is deployed by *Buddhaguhya in the late eighth century in his commen- 
tary on the passage of the Mahavairocanabhisambodhi cited above. 308 He argues 
that what those who attack this Tantra for containing elements proper to the 
non-Buddhist Tantras fail to realize is that those Tantras too were taught by the 
omniscient Buddha. 309 So it follows that there nothing inherently un -Buddhist in 

purvam kalpe-m asmim savistare || 54 pascad anyo janah prahuh kalpamantram 
prthak prthak | Humburoh (corr. : tumburuh Ed.) sarthavahasya tryambakasya tu 
*dhimatah (corr. : dhimateh Ed.). 

308 *Buddhaguhya's teaching in the Kriya- and Carya- divisions of the Tantras is said 
by Gzhon nu dpal (Blue Annals, p. 351) to have been pre-eminent in Tibet dur- 
ing the first transmission of Esoteric Buddhism, from the latter half of the eighth 
century; and this is confirmed by the Tibetan inventory of Buddhist texts in trans- 
lation compiled in the Ldan dkar palace in the early ninth century. Its small sec- 
tion of Tantras (gsang sngags kyi rgyud: entries 316-328) consists of nine texts of 
this class together with commentaries on the last four, of which three are ascribed 
to our author, those on the Vairocanabhisambodhi, the Sarvadurgatiparisodhana- 
tejordjakalpa, and the Dhydnottara. The entry on the fourth commentary, that on 
the Subdhufpariprcchd], lacks the name of its author, but it is at least probable 
that it was from the same hand, since no other Indian commentary on this text is 
known. The loss of the Sanskrit originals of these and other works of early exegesis 
has left us without the means of confirming that his name, rendered Sangs rgyas 
gsang ba in Tibetan, was indeed Buddhaguhya, as modern scholarship has gener- 
ally assumed. The evidence is inconclusive. For when the name appears in Tibetan 
sources in transcription rather than translation we find sometimes Buddhaguhya 
and sometimes Buddhagupta. We see the latter in the Ldan dkar inventory (Lalou 
1953, p. 326: slob dpon Bu ddha gu pta) and both forms are found in the colophons 
of the translations of his works in the Tenjur (Hodge 1994, p. 70). The Tenjur con- 
tains a letter (Toh. 4194) in which *Buddhaguhya addresses the Tibetan emperor 
Khri srong lde btsan, who ruled from c. 756 until c. 797 (Dotson 2007) and offi- 
cially adopted Buddhism c. 779. From it we learn that he was invited to Tibet by 
Khri srong lde btsan but declined the invitation on the grounds of failing strength, 
sending instead his commentary on the Mahavairocanabhisambodhi. 

309 rNam par snang mdzad mngon par byang chub pa'i rgyud chen po'i grel, f. 158v4- 
6: de la gsang ba'i bdag po ma 'ongs pa'i dus na sems can bio zhan pa zhes pa 
nas | de dag gis sngon sems can mams la phan par dgongs pai phyir | 'di thams 
cad bstan par rob tu mi shes so zhes pai bar du lha mams kyi kha dog gang 
yin pa dkyil 'khor yang de yin par gsungs pa | dbang po dang me'i dkyil 'khor la 
sogs pa ni | 'jig Hen pa'i rgyud la yod kyi | 'jig Hen las 'das pa'i rgyud \ bya ba'i 
rgyud dang spyod pa'i rgyud kun las mi 'byung bas na \ 'di ni sangs rgyas gsungs 


Genesis and Development of Tantrism 

Buddhist Tantric practice, however closely it may resemble the Saiva; and Bud- 
dhists, therefore, once they have understood this fact, may devote themselves 
with full confidence to the rituals of the Mahavairocanabhisambodhi. 

The Sarvatathagatattvasamgraha and the First Inroads of Sakta Saivism: Pos- 
session, Goddesses, and the Sacralization of Sex 

After the time of this text Tantric Buddhism did not, as one might expect, 
rest content with the degree of assimilation of Saivism it had already achieved, 

pa ma yin no zhes zer te | gang 'jig rten gyi | rgyud mams kyang | sangs rgyas 
boom Idan 'das thams cad mkhyen pas sems can mams so so'i dad pa dang rjes 
su mthun par mi shes pa zhes pa'i phyir ro zhes pa ste 'The statement that be- 
gins "O [Vajrapani,] Lord of the Yaksas, in time to come [there will arise] people of 
inferior understanding" refers to people who do not understand all that [the Bud- 
dha] has taught for the welfare of past beings. [The Buddha] has taught [here] 
that the colour of the Mandalas should be the same as those of [their presiding] 
deities. But some will say that the Mandalas of Isvara and of fire and the rest are 
found in the mundane Tantras [of the outsiders] and not at all in the supramun- 
dane Tantras [of Buddhism, that is to say,] in the Kriyatantras or Caryatantras, 
and that therefore they were not taught by the Buddha, [doing so] because they 
do not understand that the Blessed omniscient Buddha, in conformity with the 
various faiths of living beings, also taught [these] mundane Tantras'. This doc- 
trine that all teaching is the Buddha's, that he has taught variously in the appear- 
ance of the Buddha, Siva, and others, is set out in the Mahavairocanabhisambodhi 
in a passage that survives in Sanskrit through its citation in the Ndmamantrd- 
rthdvalokini, Vilasavajra's eighth-century commentary on the Ndmasamgiti, on 
verse 42, f. 31vl-32r2: tathd coktam srivairocandbhisambodhitantre | bhaga- 
vantas tathdgatd arhantah samyaksambuddhdh sarvajhajhanam prdpya tat sarva- 
jhajhanam sarvasattvebhyo vibhajya ndndnayair ndnabhiprdyair ndnopdyanayair 
dharmam desayanti sma \ kesam cit srdvakaydnanayam kesam cit pratyekabuddha- 
yananayam kesam cin mahdydnanayam kesam cit pancdbhijhajhdnanayam kesam 
cid devopapattaye kesam cin manusyopapattaye ydvan mahoragayaksardksasd- 
suragandharvagarudakinnarddyupapattaye dharmam desayanti sma \ tatra ke 
cit sattvd buddhavaineyikd buddharupena pasyanti. ke cic chrdvakarupena ke 
cit pratyekabuddharupena ke cid bodhisattvarupena ke cin mahesvarariipena ke 
cid brahmarupena ke cin ndrdyanarupena pasyanti sma | ke cid vaisravana- 
rupena ydvan mahoragamanusydmanusyarupena pasyanti sma \ svakasvakair 
vacanoddhdrananayair vividherydpatha<m> vyavasthitam | tac ca sarvajhajhanam 
ekarasam yad uta tathatdvinirmuktirasam ity aha mahdvairocana iti. This is 
closely related to and probably derives from the vaineyadharmopadesah, the eighth 
Prakarana of the second Nirvyuha of the Kdrandavyuha (pp. 268-269). The Sad- 
dharmapundarika likewise teaches (pp. 251-252) that Avalokitesvara assumes all 
kinds of forms, including that of Siva, in order to teach living beings in considera- 
tion of their particular dispositions. Strickmann informs us (1996, p. 440, n. 28) 
that this passage is present in the Chinese translation completed by Dharmaraksa 
in A.D. 286. It is probable that it is the model of the passage in the Kdrandavyuha. 
The doctrine that the non-Buddhist teachers are a device (updyah) of the Buddha 
is also taught in the fourth chapter of the Bodhisattvagocaropdyavisayavikurvana- 
nirdesasutra, which survives in two Chinese translations, the first by Gunabhadra 
in the fifth century; see ZlMMERMANN 2000, p. 18. 


The Saiva Age 

working only to infuse the new liturgical system with ever more clearly Buddhist 
purpose and meaning. On the contrary, with the Sarvatathagatatattvasamgraha, 
the next major Tantra, which was considered to be the foundational text of the 
Yogatantra class, which follows the Caryatantras in the ascending hierarchy of 
the classification of the Mantranaya, and was in existence in a shorter version by 
the end of the seventh century and expanded in the course of the eighth, 310 we 
find the beginning of a process of assimilation of Sakta Saiva language, practices, 
iconography, and concepts that would become ever more comprehensive through- 
out the rest of the Mantranaya's creativity Here we find for the first time the 
requirement that candidates enter a state of possession (avesah) at the time of 
their initiation. This feature, which is altogether alien to antecedent Buddhism, 
is the hallmark of initiation in the Saiva Kaula systems, setting them apart from 
all others. 311 The Vajracarya puts the candidate into a state of possession, has 

310 See Matsunaga 1978, pp. xvii-xvii. 

311 See, e.g., Tantrdloka 29.186c-220; Tantrdlokaviveka introducing 29.201c-202b: 
samdvesah sarvasdstresv avigdnenoktah; SANDERSON 1985, pp. 200—202; 1986, p. 
169 and fn. 2; and Wallis 2008. The centrality of possession in the Sakta Saiva do- 
main may derive from its Kapalika antecedents, since the Saiddhantika Saivas re- 
port that the Kapalikas [of the Atimarga] defined liberation as arising from a state 
of possession (avesah) by the qualities of the deity, analogous to the state of one 
who is possessed by a Bhuta (bhutdvistapurusavat [Naresvarapariksdprakdsa on 
1.61]); see, e.g., Pauskarabhdsya, p. 232: svayam dvisyate siddhah purusas tu gra- 
hair iva \ ittham caiva tu kdpdlds tat samyam muktim ucire; and Saivaparibhdsd, 
p. 156, 11. 22-24: kdpdlikdh samdvesena samyam upagacchanti | tathd hi yathd 
grahdh purusam avisanti tathesvaragund muktesv avisanti. They are distinguished 
in this context from the two other Atimargic traditions, those of the Pancarthika 
Pasupatas, who defined liberation as the transference of the state of equality with 
Siva in the manner in which one lamp is lit from another (sdmyasamkrdntivddah), 
and the Lakulas, who defined it as the arising of this state (sdmyotpattivddah); see 
Sanderson 2006, pp. 179-181. This hypothesis is strengthened by the fact that 
possession by the deity as the goal of practice is a marked feature of the Picumata 
and Yoginisamcdra of the Vidyapitha, texts in which the perpetuation within the 
Mantramarga of the Kapalika tradition of the Atimarga is particularly clear. Both 
describe the goal of their Kapalika-style asceticism as the entry of the deity pro- 
pitiated into the person of the propitiator. Picumata f. 101vl-3 (2.114c-117): 
duscaram devagandharvais tvayd cirnam mahdvratam || 115 varam varepsitam 
vatsa udyatam tu bravThi me | yadi tusto 'si bhagavan pravisa mama vigraham || 
116 vaktram prasarayasveti pravisya bhagavan prabhuh \ hrdaye bhairavo devo 
guhyaka tu gale sihitdh || 117 mataro hy ahga-m-ahgesu yoginyo sandhisu sthitdh \ 
sakinyo romakupesu putanddya tathaiva ca '[Bhairava says:l You have [nowl com- 
pleted the observance of the [Kapalika] Mahavrata, which is hard [even] for the 
gods and Gandharvas. Choose whatever boon you desire. Tell me without hesita- 
tion [what it is]. [The Sadhaka replies:] If you are pleased, O Lord, enter my body. 
Telling him to open his mouth the Lord God Bhairava enters his heart. [His prin- 
cipal Saktis,] the [four] Guhyakas occupy his neck, the Mother goddesses his limbs, 
the Yoginls his joints, and the Sakinis, Putanas, and others the pores of his skin'; 
cf. f. 335rl-2 (87.126c— 128b): bhairavasya mahdmudrd mudrdsdnaidhyakdrikd || 
127 prayuktd tu yadd mudrd laksanena vardnane \ bhdvdtmakavidhdnena sadyo 


Genesis and Development of Tantrism 

him cast a flower on to the Mandala to determine from the section on which it 
falls the Mantra-deity from which he will obtain Siddhi, and then, while he is 
still in this state, removes his blindfold to reveal the Mandala. He then conse- 
crates him with scented water from a Mantra-empowered vase, places a Vajra 
in his hand, and gives him his initiation-name (vajranama). 312 The immediate 
effects of the possession are described as follows: 313 

As soon as he becomes possessed supernatural knowledge arises [in him]. 
Through this knowledge he understands the thoughts of others; he knows all 
matters past, future and present; his heart becomes firm in the teachings of 
the Tathagatas; all his sufferings cease; he is free from all dangers; no being 
can kill him; all the Tathagatas enter-and-empower him; all Siddhis approach 
him; unprecedented joys arise [in him], causing spontaneous delight, pleasure, 
and happiness. In some these joys give rise to meditation-states, in some to [the 
mastery of] Dharams, in some to the fulfilment of every hope, and in some to the 
state of identity with all the Tathagatas. 


mantro vijrmbhati || 128 karoti sddhakdvesam japadhydnavivarjitd 'O fair-faced 
one, the Mahamudra of Bhairava draws every Mudra nigh. When it is employed 
correctly with full subjective immersion the [deity of the] Mantra immediately be- 
comes manifest. [The Mudra] brings about possession in the Sadhaka without 
[need of] Mantra-repetition or visualization'. The Yoginlsamcdra requires any- 
one who has gone through its initiation ceremony and then received consecra- 
tion (abhisekah) to adopt one of three forms of ascetic observance in order to gain 
mastery over the Vidya (vidydvratam): the Bhairavavrata, the Camundavrata, or 
the Trisastikulavrata, the observance of the sixty-three families [of the Mothers], 
which it also calls the Kapalavrata, i.e. the Kapalika. At the end of the obser- 
vance, we are told, the Mothers will enter his body: dvitiyam tu vratam vaksye 
ghoram kdpdlarupinam || 8.41 sire kapdlamukutam siramdldvibhusitam \ kare 
karnau tathd pddau asthikhandair vibhusitam || 8.42 vdme kapdlam khatvdhgam 
tathd vai daksine kare | smasdne vicaren mauni trisasti divasdni tu | 8.43 vratdnte 
tu vardrohe sarire mdtaro dhruvam | visante devadevesi dadante siddhim uttamdm 
'[Now] I shall teach [you] a second observance, the grim Kapalavrata. He should 
have a skull-crown on his head and be adorned with a garland of heads. His hands, 
ears, and feet should be adorned with pieces of bone. In his left hand he should hold 
a skull-bowl and in his right a skull-staff. He should wander in silence in a crema- 
tion ground for sixty-three days. It is certain that at the end of this observance the 
Mothers, O fair-hipped empress of the gods, enter his body and bestow the highest 

Sarvatathdgatatattvasamgraha, sections 224—234. 

Sarvatathdgatatattvasamgraha, section 226: dvistamdtrasya divyam jhdnam 
utpadyate | tena jhdnena paracittdny avabudhyati sarvakdrydni cdtitdndgata- 
vartamdndni jdndti hrdayam cdsya drdhibhavati sarvatathdgatasdsane sarva- 
duhkhdni cdsya pranasyanti sarvabhayavigatas ca bhavaty avadhyah sarva- 
sattvesu sarvatathdgatds cddhitisthanti sarvasiddhayas cdsydbhimukhibhavanti 
apurvdni cdsydkdranaharsaratipritikardni sukhdny utpadyante | taih sukhaih 
kesdm cit samddhayo nispadyante kesdm cid dhdranyah kesdm cit sarvdsd- 
paripurayo ydvat kesdm cit sarvatathdgatatvam api nispadyata iti. 


The Saiva Age 

and, after the bindfold has been removed: 314 

As soon as he sees the Great Mandala he is entered-and-empowered by all the 
Tathagatas and Vajrasattva dwells in his heart. He sees various visions of 
orbs of light and miraculous transformations. Because he has been entered- 
and-empowered by all the Tathagatas sometimes the Lord Vajradhara or the 
Buddha appears to him in his true form. From that time forth he attains all his 
goals, every desire of his mind, all Siddhis, up to the state of Vajradhara or the 

Anandagarbha gives a detailed account of the means by which the candidate 
is put into this state of possession in the Sarvavajrodaya, his manual on the rites 
of initiation into the Mandala of this Tantra, and makes it clear that entering this 
state is, as in the Kaula parallel, an absolute requirement. If the candidate fails 
to enter it by the standard means, the Vajracarya is to perform a rite to remove 
the sins that are assumed to be the cause, and if the candidate still fails to enter 
the possession state, he may not proceed further: 315 

If possession does not occur, because [the candidate] has committed [too] many 
sins, he should proceed to destroy those sins by repeatedly making the Sin- 
Destruction Mudra. With concentrated mind he should kindle a fire with sticks 
of sweet wood and burn all his sins by casting into it oblations of sesame seeds 
with the Mantra om sarvapapadahanavajraya svaha. He should make a 
simulacrum of those sins with black sesame seeds on the palm of his right hand 
and visualizing the [wrathful] syllable HUM in the centre he should offer it into 
the fire with his index finger and thumb. Then he should imagine that the sin 
is being incinerated in his body by Vajras wrapped in flames emerging from the 
fire-pit. [The candidate] will definitely become possessed. If possession does not 
occur even so, then he must not give him the consecration. 316 

314 Sarvatathdgatatattvasamgraha, section 231: mahdmandale ca drstamdtre sarva- 
tathdgatair adhisthyate vajrasattvas cdsya hrdaye tisthati \ ndnddydni ca rasmi- 
mandaladarsanddlni prdtihdryavikurvitdni pasyati | sarvatathdgatddhisthitatvdt 
kadd cid bhagavdn mahdvajradharah svarupena darsanam daddti tathdgato veti 
| tatah prabhrti sarvdrthdh sarvamanobhirucitakdrydni sarvasiddhir ydvad vajra- 
dharatvam api tathdgatatvam veti. 

315 Sarvavajrodaya, f. 61r4-vl (exposures 009a and 008b): atha pdpabahutvdd 
dveso na bhavati punah pdpasphotanamudrayd tasya punah punah pdpdni spho- 
tavydni \ samidbhir madhurair agnim prajvdlya susamdhitah \ nirdahet sarva- 
pdpdni tilahomena tasya tu || OM SARVAPAPADAHANAVAJRAYA SVAHA iti \ daksina- 
hastatale krsnatilaih pdpapratikrtim krtvd humkdramadhyam vicintya tarjany- 
ahgusthdbhydm homam kurydt | tato homakunddn nirgatya jvdldmdldkulair va- 
jrais tasya sarire pdpam dahyamdnam cintayen niyatam dvisati | evam api 
yasydveso na bhavati tasydbhisekam na kurydd iti. 

316 Cf Tantrdloka 29.29.210-211b: athavd kasyacin naivam dvesas tad dahed imam 
| bahir antas coktasaktyd pated ittham sa bhutale || yasya tv evam api sydn na 
tarn atropalavat tyajet 'Or, if some rare person does not become possessed by this 


Genesis and Development of Tantrism 

It is certain that the possession intended is not nominal or figurative. For 
Anandagarbha tells us that once the Vajracarya is sure that the candidate is in 
this state he should use him as an oracle: 317 

Then when the Acarya has ascertained that [the candidate] is possessed he should 
form the Samayamudra of Vajrasattva and address him with [the Mantras] HE 


SATTVA nrtya vajra (Dance, O Sattva; Dance, O Vajra). If he is indeed 
possessed he will adopt the Vajrasattvamudra. Then the Acarya should show the 
Mudra of the Vajra Fist. By this means all the deities beginning with Vajrasattva 
make themselves present [in him]. Then he should ask him something that he 
wishes [to ascertain], with the following [procedure]. He should visualize a Vajra 
on the tongue of the possessed and say Speak, O Vajra. [The candidate] then 
tells him everything [that he wishes to know]. 318 

means he should visualize him being burned both internally and externally by the 
Power [of the Mantra] taught above. By this means he will fall to the ground. 
If a person does not achieve [the state of possession] even by this means then 
in this [system] he must cast him aside like a stone'. Falling to the ground is 
commonly mentioned in Kaula texts as the consequence of initiatory possession; 
see, e.g., Matasdra f. 39v2-3: ydvanmdtram vihvalam ca vedhayet pdsapahjaram 
| pdsastobhdt pataty dsu bhutale ndtra samsayah; Jayadrathaydmala, Satka 
4, bhumikdpatalah, f. 191v (v. 105ab): saktiksobhdt tadd 
yogi viddho patati bhutale; Devidvyardhasatikd f. 16v: 197 tatksandt patate 
bhumau chinnamula iva drumah; Chummdsamketaprakdsa, first surviving verse: 
[ta]ddrkpdtamahodaydt | bhumau sampatitah ksiprdc chinnamula iva drumah; 
Urmikauldrnava f. 9r3: *pahcdvasthdgatah (em. : pancdvasvagatah Cod.) sdksdt 
sa viddhah patate bhuvi; f. 19v5-6 (2.230-231): pracalanti *mahdpdsd (corr. : 
mahdpdsam Cod.) dvesam tasya jdyate \ dnando hy udbhavah kampo nidrd ghurmis 
tu pancami || tattvaviddhasya devesi pancdvasthd bhavanti hi | sa viddhah patate 
bhumau vajrapdtdd ivdcala<h>; the Kaula Vrddhasvacchanda ff. 17v24-18r2, Ed. 
10.15c— 17a (using this MS alone): jhdtvd srisaktisamkrdmam sadevdsuramdnusdn 
|| *vedhayen (em. : vedayen Cod. Ed.) ndtra sandeham pdtayet parvatdny api || 
*sakrtsamkrdmayogena (Cod. : cakrdt sahkrdmayogena Ed.) '^chinnamula (Ed. : 
chinnamulam Cod.) iva drumah || patanti dehinah sarve; 10.25ab, Ed. 10.25ab: sa 
viddhah patate bhumau *vajrdghdtdd ivdcalah (em. : vajrdghdtam ivdcalam Cod. 

317 Sarvavajrodaya, f. 61v2— 3: tatah samdvistam jndtvdcdryena HE VAJRASATTVA HE 
baddhvoccdranlyam | punar NRTYA SATTVA NRTYA VAJRA iti | sa ced dvistah 
srlvajrasattvamudrdm badhniydt | taddcdryena *vajramustimudropadarsaniyd 
(niyd corr. : niydh Cod.) | evam sarve srivajrasattvddayah *sdnnidhyam (corr. 
: sannidhyah Cod.) kalpayanti | tato 'bhipretavastu prcched anena \ jihvdydm 
*tasydvistasya (em. : tasydvistasydvistasya Cod.) vajram vicintya bruhi vajra iti 
vaktavyam | tatah sarvam vadati. 

318 The inducing of possession in persons so that they may be used as oracles, is not 
restricted in Tantric Buddhism to the context of initiation. It is also seen as an 
independent procedure in which the medium is a young boy or girl. We find it in 
the Tantra Subdhupariprcchd in a section partly translated and partly paraphrased 
from the Chinese by Strickmann (1996, pp. 222-226), a work that was translated 


The Saiva Age 

into Chinese (T. 895) by Subhakarasimha in 726 and was in the hands of the Chinese 
monk Wu-xing in 674 (Hodge 2003, p. 18). We also see it in the Suji li yan mo xi 
shou luo dian shuojia lu luo a wei she fa 'The quickly effective method of possession 
(avesah) taught by the god Mahesvara' (T. 1277). This short scriptural text, whose 
translation from the Sanskrit is assigned to Bukong (Amoghavajra) and to a date 
between 746 and 774, claims in its preamble that it is a teaching given by Siva 
(Mahesvara) to Narayana on Mt. Gandhamadana in answer to the latter's request. 
It sets forth a procedure to induce the messenger (Duta) of Mahesvara to possess a 
young girl aged seven or eight so that he can then use her while she is in this state 
to answer any questions he has concerning the future. He should have her fast by 
eating nothing but pure foods for three or seven days. Then on an auspicious day 
he bathes her, anoints her with unguents, gives her clean clothes, puts camphor in 
her mouth, sits facing East, smears a low wooden platform with sandalwood-paste, 
has the girl stand on it, scatters flowers in front of her, sets up a vessel of Argha 
water, takes incense, empowers it seven times with the Mahamudramantra, lights 
the incense and fumigates the girl's hands with it, takes a red flower, empowers it, 
places it in her hands, and passes his hands over her face. Then, with his hand 
forming a Mudra he touches and thus empowers five parts of his own body and then 
with the same Mudra touches the girl's head, her mouth, his heart, and his navel 
visualizing in these the symbols of fire, water, earth, and wind respectively. He then 
empowers his two legs, visualizes Garuda, puts the armour-Mantra on the girl's 
body, and visualizes himself as Mahesvara, three-eyed, with the digit of the moon 
on his crown, blue-faced, eighteen-armed, and brandishing various weapons, with 
a snake as his sacred thread, wearing the bleeding hide of an elephant. He then 
protects her with recitation, empowers flowers, incense, and Argha water with the 
Mahamudramantra, and seals the ten directions. Then facing the girl the Sadhaka 
recites the Mantra of Mahesvara's Duta. The girl will start to tremble. This reveals 
to him that the Duta has entered her. He then snaps his fingers and recites the 
Mantra. If she does not fall into the possession trance he should recite a further 
Mantra to incite the Duta to enter her. By this means the result is certain. He then 
interrogates her about good and bad in the future and is told whatever he wishes to 
know. This account is based on an oral translation of the Chinese text very kindly 
provided by my colleague Notake Miyako (Leipzig). A French translation of part 
of the text, without the visualizations, is given in Hobogirin, p. 7. 

Here too the model is Saiva, as the preamble and content of this text suggest. 
Putting children into a possession-state is already present in the earliest liter- 
ature of the Saiva Mantramarga, where we find the use of Ksatriya and brah- 
min boys for this purpose; see Nisvasatattvasamhita f. 82vl-2 (Nisvasaguhya 
10.116— 117b): athavesam kartukafmah] + + ksatrakumdrakam | sndpayitvd tarn 
ekam tu suddhadehah savasakam || purvamukham sthapayitva hy udakenavesayet; 
f. 112v6 (Nisvasaguhya 17.30): athavesam kartukamo brahmanakumarafkam + 
ujdakena snapya tenaiva tadyamanam avesayed vacayd moksah. The ritual also 
appears in narrative literature. The Kathasaritsagara (70.55-63) tells a story of 
an ash-smeared ascetic, a pupil of Suddhakirti, who has mastered many Mantras 
and claims to have done this with a Ksatriya boy (56cd: subhalaksanam asadya 
kamcit ksatrakumdrakam), who in his trance revealed the whereabouts of many 
miraculous herbs and elixirs (57: sa kumarah samavistah prsto nanavidhani me 
| siddhausadhirasaksetrany udiryedam athabravlt), and, finally, a palace of the 
Nagas in a pollen-covered pond in the jungles of the Vindhya mountains, where, 
with the help of Viras, he could obtain a sword that would make him lord of the 
Siddhas. The procedure is referred to there as a svasthavesah 'a [rite of caus- 


Genesis and Development of Tantrism 

Nor is possession restricted in the Sarvatathagatasamgraha to the context of 
initiation. The term dvesah is used repeatedly in the text to denote the state that 

ing oracular] possession in one who is healthy (svasthah) [in body and mind]' 
(70.56ab: so 'ham kadapy akaravam svasthavesam prasahgatah), and it appears 
under this name frequently in the Vidyapitha, where in accordance with that lit- 
erature's Sakta character the medium is, as in the Su ji li yan mo xi shou luo 
dian shuo jia lu luo a wei she fa, a young girl. We see this in Jayadrathayamala, 
Satka 2, f. 19r9— v3 (6.54c— 59): kanyam sulaksanopetam dhutavasam manoharam 
|| 55 svalamkrtam atah krtva ratrav eva mahesvari \ dattva dhupam tato vidyam 
avart'ye<t> sadhakesvarah || 56 tavad avartayed ghoram yavad avesam apnuyat 
| divyabhaumantariks*adyam (conj. : adya Cod.) avesam kurute ksanat || 57 
hastardham ca ksitim tyaktva tisthate vikrtanana \ tada mahalipisitais tar- 
payet suravandite || 58 prahvas ca pranato bhutva prcchet sadhakasattamah | 
sadasivadiksityante yavan manasi rocate || 59 tat sarvam kathayed devi yad anyam 
va hrdi sthitam | evam prstva visarjeta pranamya paramesvari 'Then, at night, O 
Mahesvari, the lord among Sadhakas should adorn a pretty young girl endowed 
with excellent characteristics and wearing freshly washed clothes, fumigate her 
with incense, and then begin to repeat [the Vidya of] Ghora. He should con- 
tinue to repeat it until she becomes possessed. Immediately [her understand- 
ing] penetrates all that is in the heavens, on the earth, and in the sky. With 
her face contorted she hovers half a cubit above the ground. Then, O honoured 
by the gods, he should gratify her with offerings of wine and meat. He should 
then bow low before her and put his questions to her. O goddess, she will tell 
him all that he wishes to know in the whole universe, from the level of Sadasiva 
down to Earth, and other matters that are concealed in his heart. When he 
has interrogated her in this way, O Paramesvari, he should prostrate himself in 
veneration and allow her to leave'; and Jayadrathayamala, Satka 3, f. 99v2-6 
(14.70—76): atha sadhayitum vanche<t> svasthavesanam uttamam \ tada kanyam 
samaniya sarvalaksanalaksitam \\ 71 asane tarn pratisthapya sugupte varamandire 
| raktakrsnambaradharam raktasrakkanthasobhitam \\ 72 subhasanastham tarn 
kuryat palaliparipuritam \ avyucchinnam dahed dhupam vidyam avartayet tatah 
|| 73 tada sa kampate kanya ghurnate hasate punah \ ghantam pravadayet tatra 
mahamantra*vidhau (conj. : vikai Cod.) sthitah \\ 74 tata avisate turnam devadevT 
krsodarl \ tyaktva bhumim tisthate sa tada *sa (corr. : sa Cod.) pranatah puman 
|| 75 tarpayet paramesanim nanabalyopaharatah \ tada sadhakamukhyaya vadate 
*manasepsitam (corr. : manasipsitam Cod.) || 76 bhutam bhavyam bhavisyam ca 
kalatrayam athakhilam \ brahmandodaraga varta<h> sadhakaya vadaty asau 'If 
he desires to accomplish the supreme rite of svasthdvesah he should bring a young 
girl who possesses all the necessary characteristics and set her on a seat in an excel- 
lent building that is well concealed. Her seat should be of fine quality. She should 
be dressed in a dark red garment; her neck should be adorned with a garland of 
red flowers; and her mouth should be filled with wine and meat. He should burn 
incense without interruption and then repeat the Vidya again and again. Then the 
girl begins to tremble, swoon, and laugh. Established in the procedure of the Great 
Mantra he should ring his bell. The emaciated Goddess will immediately enter [the 
girl], who will then rise and hover above the ground. The Sadhaka should then 
prostrate himself before her and gratify the Goddess with the offering of a various 
Balis. Then [speaking through the girl] she will tell that excellent Sadhaka what- 
ever he desires to know. She will explain to him [anything he wishes to ascertain 
in] the three times, past, present, and future], all events within the entire sphere of 


The Saiva Age 

the practioner must induce in himself in order to accomplish both his Siddhis and 
his enlightenment, typically in the compound vajravesah 'possession by Vajra'. 
For example: 

For by means of possession by [Vajrajsattva enlightenment will quickly be at- 
tained. 319 

When he has given rise to dvesah in this way whatever form he meditates on as 
his own will automatically become Buddha in form. 320 

When vajravesah has arisen he should visualize the water as an embodiment of 
the Vajra. Quickly achieving success he will be able to walk on [that] water. 321 

Once he has generated vajravesah, if with concentrated mind he makes a slight 
clap with his palms in the Vajranjali [gesture] he can subject to his control even a 
mountain. 322 

Likewise, by virtue of the practice of dvesah, if he stretches out [his hands in] the 
Vajra gesture and strikes together the tips of his fingers he can kill a hundred 
families. 323 

Two other features of this seminal text evidence the influence of Sakta 
Saivism. The first is the fact that after teaching the Vajradhatumandala in 
its opening section it goes on to teach the Vajraguhyamandala, in which the 
five Tathagatas are replaced by goddesses: Vairocana at the centre by Va- 
jradhatvisvari and, around her in the four directions, Aksobhya by Vajravajrini, 
Ratnasambhava by Ratnavajrini, Amitayus by Dharmavajrini, and Amogha- 
siddhi by Karmavajrini. 324 In the preamble Vajrapani makes the following 
joyous declaration (udanam): 325 

Ah, how benevolent is the Bodhicitta to all beings! For the Buddhas take on even 
female form to accord with [the expectations of] their disciples (vineyavasat). 

319 Sarvatathagatatattvasamgraha, section 167: yat sattvavesayogad dhi ksipram bo- 
dhir avapyata iti. 

320 Section 238: tathaivdvesam utpddya yad rupam svayam dtmanah | *bhdvayen (em. 
: bhdvayan Ed.) bhavate tat tu buddharupam api svayam. 

321 Section 238: vajrdvese samutpanne vajrabimbamayam jalam | bhdvayet 
ksiprasiddhas tu jalasyopari cahkramet. 

322 Section 247: vajrdvesam samutpadya talam dadyat samahitah | vajranjalitalaih 
suksmam parvato 'pi vasam nayet. 

323 Section 247: tathaivavesavidhina vajrabandhe (conj. : bandha Ed.) prasarite 
agrdhgulisamasphotad dhanet kulasatam ksanat. 

324 Sarvatathagatatattvasamgraha, sections 319—327. 

325 Sarvatathagatatattvasamgraha, section 322: aho hi bodhicittasya sarvasattva- 
hitaisita \ yad vineyavasad virah strirupam api kurvate. 


Genesis and Development of Tantrism 

The second is the incorporation of sexual intercourse into the activities of 
worship as a higher form of practice. This element is not conspicuous because it 
is not mentioned in the treatments of the principal Mandalas taught in the text 
and it was therefore easily pushed out of view when this text was propagated 
in China and thence in Japan. It is present nonetheless as an esoteric teach- 
ing reiterated many times throughout the text in the form of passages teaching 
that the pleasure of sexual union and indeed other sensual delights are a means 
both of worshipping the Buddha and of attaining Siddhis when combined with 
meditation on one's Buddha nature. For example: 

1: If after generating a firm intention to attain enlightenment he meditates on 
himself as the Buddha and worships himself [as the Buddha] with the pleasure of 
sexual intercourse he will obtain the joys of the Buddha himself. 

2: He will quickly become equal to Vajrasattva if he presents the pleasures of em- 
bracing the body of any [womanl as offerings to the Buddhas. He will be become 
equal to Vajraratna if he presents the pleasures of grasping [her] hair in intensely 
felt love as offerings to the Buddhas. He will become equal to Vajradharma if he 
presents the exquisite pleasures of kissing while immersed in intense sensual 
delight as offerings to the Buddhas. He will become the equal of Vajrakarma if 
during his worship he completely offers up to the Buddhas the pleasures of the 
union of the two sex organs. 

3: He will attain success in the Mandala by means of the union of the two sex or- 
gans while meditating with fully concentrated mind on the meditation state that 
embodies all things. 

4: Non-detachment from sensual pleasures: this is the greatest and purest rule 
of discipline [for an initiate] in the family of the Tathagatas. It may not be trans- 
gressed even by the Buddhas. 

5: There is no religious duty purer than [the exercise of] sexual desire, the be- 
stower of all joys. This, which brings about Siddhi, is the highest duty in the 
family of the Tathagatas. 

6: During worship with the four prostrations he will quickly attain Siddhi if when 
exhausted from the exertion of love-making he offers [to the Buddhas] the plea- 
sure which that love-making aroused. 

7: He will attain Siddhi if while meditating with in-turned mind on the purity of 
lust he worships the Buddhas with the drops of his semen. 326 

326 1 Section 288: bodhicittadrdhotpadad buddho 'ham iti cintayan | ratya tu puja- 


The Saiva Age 

The Guhyasamdja: copulating deities, sexual initiation rites, and the sacraliza- 
tion of impurity 

In the next phase of the Mantranaya, seen in the Guhyasamdja, also a 
product of the eighth century, this esoteric eroticism has moved to the fore- 
ground; and this is apparent from the very beginning of the text. For the place 
where the Buddha is said to have been residing at the time that he revealed this 
Tantra, which was expected to be stated in the preamble (nidanavakyam) of any 
scripture claiming to be Buddhist, is not one of the familiar sites of revelation 
such as Rajagrha, Dhanyakataka, or, as in the Sarvatathdgatatattvasamgraha, 
the Akanistha heaven, but the vaginas of the goddesses Locana, Mamaki, 
Pandaravasini, and Tara, that is to say, a timeless, unlocated bliss: 327 

[I aver that] I once heard the following [teaching]. The Venerable Lord was re- 
siding in the vaginas of the Vajra-women of the body, speech, and mind of all the 
Tathagatas . . . 

and this surprising relocation, no doubt provocatively shocking in its time, 
became standard in the subsequent literature of the Mantranaya, both in 
texts closely related to the Guhyasamdja and in the next wave of texts, the 
Yoginitantras, in which the influence of the Sakta Saiva tradition became much 
more intense and pervasive. 328 

yann dtmd labhed buddhasukhdny api; 2 Sections 549-553: sarvakdyaparisvahga- 
sukhapujdh svayambhuvdm | nirydtayan bhavec chighram vajrasattvasamo hi sah 
|| drdhdnurdgasamyogakacagrahasukhdni tu | niryatayams tu buddhanam va- 
jraratnasamo bhavet \\ drdhapritisukhdsakticumbitdgryasukhdni tu | niryatayams 
tu buddhanam vajradharmasamo bhavet \\ dvayendriyasamdpattiyogasaukhydni 
sarvatah \ niryatayams tu pujdydm vajrakarmasamo bhaved iti; 3 Section 1825: 
visvarupasamddhim tu bhdvayan susamdhitah | dvayendriyasamdpattyd mandate 
tu sa sidhyati; 4 Section 2168: kdmdndm avirdgas tu samayah sumahdn ayam \ 
tathdgatakule suddho ndtikramyo jinair api; 5 Section 2175: rdgdc chuddhataro 
ndsti dharmah sarvasukhapradah \ tathdgatakule 'py esa dharmah siddhikarah 
parah; 6 Section 2506: suratasramakhinnas tu tat saukhyam suratodbhavam 
catuhprandmapujdydm nirydtya laghu sidhyati; and 7 Section 2651: antargatena 
manasd kdmasuddhim tu bhdvayan | svaretobindubhir buddhdn pujayan siddhim 
dpnuydt. Other passages advocating sexual intercourse in worship are to be found 
in sections 475-479, 525-529, 929-932, 1184, 1790-1792, 1918-21, 2071-2074, 
2158-2159, 2177, 2360-2363, 2415-2416, 2419-2421, 2425, 2439, 2443, 2445, 2504, 
2508, 2510, 2512, 2516, 2672, 2720, 2950, and 2951. 

327 Guhyasamdja, preamble: evam mayd srutam ekasmin samaye | bhagavdn sarvata- 
thdgatakdyavdkcittahrdayavajrayosidbhagesu vijahdra. 

328 This same formula, or a variant, is seen in the Vajramdld (rDo rje phreng ba), f. 
208r2-3: bcom Idan 'das de bzhin gshegs pa thams cad kyi sku dang gsung dang 
thugs kyi sning po rdo rje btsun mo'i bha ga mams la (as in the Guhyasamdja]), the 
Krsnayamdri (sarvatathdgatakdyavdkcittasarvavajrayosidbhagesu), and in those of 
the Yoginitantras that have a nidanavakyam: the Hevajra and Samputodbhava 
(both as in the Guhyasamdja), the Vajrdmrta (f. lvl: sarvatathdgatakdyavdk- 
cittahrdayavajrdmrtaguhyapadmesu), Vajrdrali (rDo rje a ra li, f. 171r2-3: de 


Genesis and Development of Tantrism 

In the Guhyasamdja the male deities, now multi-faced and multi-armed in 
a fusion of Saiva and Buddhist iconography, are represented and visualized cop- 
ulating with their consorts; 329 and both initiation and subsequent practice now 
involve copulation with a female partner, as in the Saktism of the Saivas. 330 A 
further borrowing from the Vidyapitha is evident in the introduction of a cru- 
cial element of what that tradition calls 'non-dualistic practice' (advaitacarah) 
and both traditions call 'practice free of inhibition' (nihsankacarah), namely the 
offering to the deities of such 'impure' substances as urine, faeces, semen, and 
blood, and their sacramental consumption. 331 

bzhin gshegs pa thams cad kyi shes rab kyi pha rol tu phyin pa'i bha ga 
la [*sarvatathdgataprajndpdramitdbhage]), Candamahdrosana (sarvatathdgata- 
kdyavdkcittahrdayavajradhdtvlsvaribhage), Abhidhdnottara (f. Iv3: sarvatathd- 
gatavajrakrodhaddkaddkinlguhyahrdayesu), Samvarodaya (sarvatathdgatakdya- 
vdkcittavajrayoginibhagesu), and Ddkdrnava (f. lvl: mahdviresvarasarvatathd- 

' This is the case in both of the major Mandalas based on this Tantra, that of saffron- 
coloured Vajrasattva-Manjuvajra and that of black Aksobhya. For the full iconog- 
raphy of these pantheons see Nispannayogdvali A, pp. 1-7; B, pp. 1-12. The prin- 
cipal difference between them is that in the Aksobhyamandala only Aksobhya, the 
central deity (cakresvarah) and the ten wrathful Krodharajas that form the outer 
protective circle are represented embracing consorts (sasvdbhaprajndh), whereas 
in the Manjuvajramandala this is also the case with the four Tathagatas (Vairo- 
cana, Ratnasambhava, Amitabha, and Amoghasiddhi) that occupy the four direc- 
tions around the central deity. All the deities in both Mandalas are three-faced and 
six-armed and all except the Krodharajas, who stand in the aggressive Pratyalidha 
posture, are seated in the Vajraparyanka posture. None of the deities has any of the 
Kapalika attributes that mark the iconography of the Yogimtantras, namely the 
skull-bowl, skull-staff, bone-ornaments, and coating of ash. 

' The Guhyasamdja proper (chapters 1-17) gives little detail in its account of initia- 
tion and makes no mention of the involvement of a consort, speaking of the neces- 
sity of acquiring such a partner only in the context of the post-initiatory practice 
known as the vidyavratam; see 16.93: sodasdbdikdm grhya sarvdlahkdrabhusitdm 
| cdruvaktrdm visdldksim prdpya vidyavratam caret After obtaining a girl of six- 
teen with a charming face and wide eyes, adorned with every adornment, he should 
practice the Vidyavrata [with her]'. The supplementary 18th chapter, however, the 
Samdjottara, gives an account of the initiation involving copulation in its vv. 113- 

L See, e.g., Guhyasamdja 4.21: vinmutrasukraraktddin devatdndm nivedayet | evam 
tusyanti sambuddhd bodhisattvd mahdsaydh 'He should offer to the deities such 
things as urine, faeces, semen, and blood. In this way the noble Buddhas [and] 
Bodhisattvas are gratified' (cf. the following in the Guhyasamdja's satellite Tantra 
Vajrahrdaydlamkdra, Patala 3 [rDo rje snyingpo rgyangyi rgyud f. 39v3-4]: bshang 
gci khu ba khrag mams ni | dung chenpo ru bzhag byas te | lha mams la ni dbul bar 
bya 'He should place faeces, urine, semen, and blood in a human skull [mahdsahkhe] 
and offer them to the deities'); 6.21: vinmutrdhdrakrtydrtham kurydt siddhiphald- 
rthinah \ sidhyate 'nuttaram tattvam bodhicittam andvilam 'If he desires to attain 
Siddhi he should consume faeces and urine. [By this means] he will master the 
ultimate reality, the spotless Bodhicitta'; 7.33ab: samaydt ksared retam tu vidhind 
pibet phalakdnksinah 'In accordance with the rule of the discipline he should ejac- 


The Saiva Age 

ulate his semen and drink it if he desires to attain his goal'; 12.47cd: pahcdmrta- 
prayogena vajrasattvatvam dpnuydt 'By the use of the Five Nectars he will attain 
Vajrasattva-hood'; 16.7ab: avasyam eva ddtavyam vinmutrddyam visesatah 'One 
must especially offer [to the Mandala] such substances as faeces and urine'; 17.47: 
vinmutrasukraraktdndm jugupsdm naiva kdrayet | bhaksayed vidhind nityam idam 
guhyam trivajrajam 'He must not feel disgust at faeces, urine, semen, and blood. 
He must regularly consume [them] according to the rite. [For] this is secret of the 
three Vajras [of body, speech, and mind]'; 18.67c-68b: simhavad vicaren mantri nir- 
visahkena cetasd | ndkdryam vidyate hy atra ndbhaksyam vidyate tathd 'He should 
wander [fearlessly] like a lion, with a mind free of inhibition. For him there is 
nothing that he may not do, nothing that he may not eat'. On advaitdcdrdhlnih- 
sahkdcdrah and the use of such substances, the Five Nectars (pahcdmrtam), in the 
rites of the Sakta Saivas see Sanderson 2005c, pp. 110-113, fn. 63; and, e.g., Vi- 
malaprabodha, Kdlikulakramdrcana, f. 65r3— v4: atha nityanaimittikakdmydrcane 
kuladravyaganam likhyate \ paldndum lasunam grhjam lambusam lavatarkasam | 
vdmdpuspam puspabandham astau dravydni kaulike \\ sivdmbu surd raktamadyam 
mahdtailam ca sidhukam | kundagolodbhavam sukram peydny astau kuldgame 
|| matsyam mdmsam mahdgottham sthalajdkdsanirajam | mahdmdmsam mrgam 
caiva bhaksydny astau kulakrame || mdtahgi kajjalT saundl kandukl carminl 
dhvajd | chippi vesyd susambaddhd grdhyaitdh kdlikdkule || nihsahkdcdramdrgena 
pujanam ca bhaved yadi \ taddsau sidhyate *devi (em. devi Cod.) tair *bhuktvd 
bhdvitd yadi (conj. : bhuktam bhdvitam yadi Cod.) || tatpdnasparsandhdrdt 
pdsacchedakarT smrtd | *gopitam (conj. : gopitais Cod.) tan mayd purvam ad- 
vaitdcdrasobhanam . Cf. in the Mantranaya, e.g., the Sarvadevasamdgamatantra 
(lost in Sanskrit, apart from citations, and not translated into Tibetan) quoted 
in the Tattvasiddhi of Santaraksita, A f 96v3-6, B f. 39vll-13 (Tib. f 30r5-7): 
*nirvikalpena bhdvena (em. [Tib. mam par mi rtog sems kyis ni] : nirvisahkena 
bhdvena AB) sarvakarmdni sarvadd | *dcaren (conj. : dcdran B : dcdra A [Tib. 
spyod pa]) nirvisahkena tapasdm *uttamottamam (em. [Tib. mchog gi mchog] 
: uttamamstapah B : uttamdtapa A) || *visaydn sevamdnasya (em. [Tib. yul 
mams *bsten (corr. : bston Cod.) par gyur pa na] : visaydhgavimdnasya AB) 
nirvikalpena cetasd | *kutsddhikam na va cet tat (tentative conj. [cf. Tib. smod 
par gyur pas mi gnod pa] : kutsddhikam na va cetas B : kutsddhikandceta A) 
tat tapo *duratikramam (corr. : duratikramah AB) || yas tu sarvdni karmdni 
*prajhayd (em. [Tib. shes rab kyis] : prajhdyd B : prajhdydyd A) viniyojayet | 
*sd ca sunyapade yojyd (em. [Tib. de yang stong pa'i gnas su sbyar] : sarvdh 
siinyapade yojya B : sarvasunyapade yojya A) *tapo (em. [Tib. dka' thub] : tathd 
AB) hy esa mahdtmandm || *prajndsamkrdntirupena (B [Tib. shes rab pho ba ngos 
pos ni] : prajhdsamkrdtirupana A) nirvikalpena cetasd \ *nihsahkdcdrasamcdras 
(em. [Tib. dgos pa med par kun spyod] : nihsahkdndrasancdrahs AB) *tapas 
tesdm (B [Tib. de'i dka' thub yin] : tapatapatesdm A) mahdtmandm. A version 
of this passage is contained in the Vajraddka, f. 3v2-4 (1.57c-62b): sopdya<m> 
sarvakarmdni nirvisahkas cared yadd || 1.58 nirvikalpena bhdvena vratdndm ut- 
tamotta*mam (em. : mah Cod.) | nirvikalpena bhdvena sarvakarmdni sarvadd 
|| 1.59 dcare<n> nirvisahkena tat tesdm *uttamam tapah (conj. : uttamdttatah 
Cod.) | visayan * sevamdnasya (em. : sevyamdnayo Cod.) nirvisahkena cetasd 
|| 1.60 *kesondukdnubhdvena (em. : kesondukasvabhdvena Cod.) tat *tapo (em. 
: tayo Cod.) duratikra*mam (corr. : mah Cod.) | yas tu sarvdni karmdni 
prajhayd viniyojayet || 1.61 sd ca sunyapade yojya tapo hy etat mahdtmandm || 
prajhd* samkrdntarupdndm (conj. : samkdsarupdni Cod.) nirvikalpena cetasd || 1.62 
nihsahkdcdra*samcdras (corr. : samcdrahs Cod.) tapas tesdm *mahdtmandm (corr. 


Genesis and Development of Tantrism 

That Tantric Buddhists possessed the specialized knowledge of the Saiva 
Mantramarga that would enable them to draw at will on the Saiva Tantras in 
this period is placed beyond doubt by an early exegetical work in the tradition 
of the Guhyasamaja. For this, the Guhyasiddhi of Padmavajra, written in all 
probability in the eighth century 332 assumes that any initiate in the practice of 
this Tantra is not only familiar with the Saiva scriptures but is able to enact 
their rituals by assuming the role of a Saiva Guru, implying thereby that such 
initiates were typically converts from the Mantramarga with experience both of 
its texts and of its practices. For it tells the adept of this tradition that in or- 
der to acquire the female consort required for his post-initiatory observance he 
should enter the home of a family of untouchables who are observant devotees 
of Siva, reveal to them one of the Saiddhantika scriptures — the text specifically 
mentions the Kdlottara and the Nisvasa — give them Mandala initiation [follow- 
ing this scripture], and then return to them the daksina that they will give him, 
taking a girl from them in its place: 333 

He should wander in other lands, in which he is known nowhere. With firm re- 
solve the Sadhaka should enter among untouchables who are devotees of Siva 

: mahdtmanah Cod.). 

332 Portions of the Guhyasiddhi have been quoted in the Carydmeldpakapradipa 
of Aryadeva: Carydmeldpakapradipa, pip. 71—72 (imam evdrtham dyotayann aha 
sriguhyasiddhau:) = Guhyasiddhi 3.71-81, 17.38; p. 77 = 6.2-3; and p. 97 = 6.45- 
49. TOMABECHI (2008, p. 175) has shown that Aryadeva's work is likely to have 
been written in the early years of the ninth century. 

333 Guhyasiddhi 8.8c-16b: paryated *anyadesesu (conj. [cf. 8.2cd: pravisya 
cdnyadesesu] : divyadesesu Ed. [Tib. bzang po'i yul du 'khyam par bya]) yatra na 
jhdyate kvacit || 9 pravisya * canty ajdtindm madhye (em. [Tib. mthar skyes nang 
du 'jug par bya]: canty ajddlndm madhye Ed.) ye tripurdntake | bhaktd jdnanti 
naivdnyam daivatam paramdrthatah || 10 *siddhdntabhdvitd nityam (em. [Tib. 
rtag tu rang gi grub mtha' bsgom (*svasiddhdntabhdvakd nityam)] :siddhyante 
bhdvitd nityam Ed.) sndnadev arcane ratdh \ kimcidaksaramdrgena *prasaktdh 
(conj. \prasakte Ed.) sdstradarsane || 11 evam pravisya tanmadhye sddhako 
drdhaniscayah | canddlaganarupena bhdvayan bodhim uttamdm || 12 *darsayec 
ca tatas tesdm dharmam siddhdntapurvakam (em. [cf. Tib. chos dang grub 
mtha' sngon 'gro ba | de nas de la ston par byed] :darsayec ca tatas tesdm 
dharmasiddhdntapurvakam Ed.) | kdlottarddi* samsiddham (em. : samsuddham 
Ed.) no cen nihsvdsasambhavam || 13 pdtayitum ca visvdse sarvdms tarns 
tantracoditdn | krtvd caivdtmanah sisydn dlksdmandalapurvakam || 14 tato yat 
samcitam dravyam tair dattam gurupujane \ tat tesdm arpayitvd tu purvam vit- 
tena samyutam \\ 15 grhitvd kanyakdm tesdm cdruvaktrdm sulocandm \ tarn 
krtvd mantrasadbhdvdbhijndm samayasammatdm \\ 16 cared vidydvratam dhimdn 
buddhatvakrtaniscayah. I have emended antyajddindm to antyajdtindm with the 
support of the Tibetan because the -ddi- is inapposite: in 8.7 the Sadhaka is 
told to enter the home of an untouchable (antyajdlayah); and in 8.1 he is told 
that it is an untouchable girl (antyajd) that he is to acquire. I take dharmam 
siddhdntapurvakam in 8.12c to mean 'dharmam preceded by [the wordl siddhdnta- 
', i.e. siddhdntadharmam, an example of a not uncommon style of periphrasis. 


The Saiva Age 

and recognize no other deity as absolute, who are inspired by the Siddhanta, al- 
ways attached to [the rituals of] bathing and deity-worship, and dedicated to the 
doctrines of its scriptures through some slight degree of literacy. After entering 
among them in the guise of an untouchable votary (canddlaganah), he should, 
while cultivating insight into the highest wisdom, instruct them in the religion of 
the Siddhanta established in such scriptures as the Kdlottara, or the Nisvdsaf 34 
and in order to win their trust he should take as his disciples all those who are 
enjoined by the Tantra after [initiating them before] the Initiation Mandala [of 
Siva]. Then he should give back to them all the goods and money that they will 
previously have gathered and given him as their offering to their Guru and take 
[instead] a girl of theirs with a beautiful face and eyes. After acquainting her 
with the essence of the Mantras and making her adhere to the rules of an initiate 
that wise one should practice the Vidya observance [with her], after resolving to 
become a Buddha. 335 

This is indeed troubling evidence for those who may be reluctant to accept that 
Buddhists would have had the familiarity with Tantric Saivism that my thesis 
of the development of the Mantranaya presupposes. 

The Sarvabuddhasamayogadakinijalasamvara: Heruka and his Yoginis, 
Kapalika iconography, the Ganamandala, and the beginning of Saiva-Buddhist 

With the Sarvabuddhasamayogadakinijalasamvara, another product of 
this century, 336 we see the beginning of the final phase of saktization. It is still 
rooted in the liturgical tradition of the Yogatantras, 337 as can be seen in the 

334 Literally "that which has arisen from the outbreath (nihsvdsah/nisvdsah) [of Siva]". 
Both forms of the name of this scripture, Nisvdsa and Nihsvdsa, are attested. 

335 Padmavajra is elaborating on Guhyasamdja 16.93: sodasdbdikdm grhya 
sarvdlahkdrabhusitdm \ cdruvaktrdm visdldksim prdpya vidydvratam caret 'He 
should take a girl of sixteen with a beautiful face and wide eyes, adorned with every 
ornament, and practice the Vidya observance with her'. 

336 It was translated into Tibetan towards the end of the eighth century or early in 
the ninth, and Amoghavajra (705-774) names it and provides a brief summary of 
its teachings in his Jin-gang-ding-jing yu-jia shi-ba-hui zhi-gui, Jap. Kongd-chd- 
gyd yuga juhatte shiiki (T 869) Key Points of the Eighteen Assemblies of the Yoga 
of the Vajrasekharasutra; see TOMABECHI 2007, p. 905. He composed this work in 
Chinese at some time between 746 and and his death in 774, but we can be sure 
that the text existed in some form, perhaps in an early stage of its development, by 
c. 740, since his knowledge of it must have been gained between 741 and 746, when 
he was in Ceylon and perhaps India gathering the Tantric literature whose analysis 
and translation into Chinese occupied the rest of his life. 

337 It is referred to by Aryadeva as a Mahayogatantra in his Carydmeldpakapradlpa, 
p. 82: adhund prapahcatdcaryd srTsarvabuddhasamdgamayogaddkinijdlasamvara- 
mahdyogatantrdd avatdryate. This term serves to distinguish it from the Yo- 
gatantras, namely the Sarvatathdgatatattvasamgraha and its satellites and to 


Genesis and Development of Tantrism 

group it with the Guhyasamaja and related texts, though which of the Yo- 
gatantras in the broad sense qualified to be considered Mahayogatantras might 
be the subject of divergence of opinion. Dlparikarasrljnana defines this class 
(rnal 'byor chen po'i rgyud) as comprising the Guhyasamaja and its explana- 
tory Tantras (vyakhyatantrani), which he lists as the Guhyendutilaka, the 
Krsnayamari, the Paramddya, the Sarvadevasamagama, the Sarvarahasya, the 
Vinayamoghafsiddhi], the Vajrajnanasamuccaya, the Vairocanamaydjala, the 
Laghukhasama , the Advaya[samata]vijay a, and the Vajrasekhara (Byang chub lam 
gyi sgron ma dka' 'grel, p. 286: de la rnal 'byor chen po'i rgyud ni dpal gsang ba 
'dus par bshad rgyud dang bcas pa dang zla gsang thig le dang gshin rje'i gshed 
nag po dang mchog dang po dang lha thams cad 'dus pa dang thams cad gsang ba 
dang 'dul ba don yod pa dang ye shes rdo kun las btus pa dang mam par snang 
mdzad sgyu 'phrul dang nam mkha' dang mnyam pa chung ngu dang gnyis medpa 
mam par rgyal ba'i rgyud dang rdo rje gtsug tor rgyud la sogs pa rgyud sde stong 
phrag bcu gnyis te rgyas par by as na grangs pa med do.) An alternative terminology 
distinguishes these more esoteric Yogatantras as Yogottaratantras, perhaps origi- 
nally in the meaning 'Supplementary Tantras (uttaratantrani) of the Yoga [class]', 
and refers to the Yogimtantras as Yoganiruttaratantras, giving the ascending series 
Kriyatantra, Caryatantra, Yogatantra, Yogottaratantra, and Yoginiruttaratantra; 
see, e.g., Ramapala, Sekanirdesapanjika, introducing verse 1, describing his teacher 
Maitreyanatha (Advayavajra) as an unsurpassed master of all of these: iha maha- 
panditavadhutasrimaitreyanathah kriyacaryayogayogottarayoganiruttaratantresv 
anuttaraguruh; Ratnakarasanti, Muktavali, p. 223, on Hevajra 2.8.10: sarvam 
iti pancavidham: kriydcaryayoga*yogottarayoganiruttarabhedena (yogottara corr. 
[=Cod., f. 45v6] : yogantara Ed.); Kanha, Yogaratnamala, p. 156 (on Heva- 
jra 2.8.10): sarvamantranayam iti pancavidham kriyacaryayogayogottarayoga- 
niruttarabhedena; Advayavajra, Gudhapada, f. 6r6— 7: vajram pancajnanatmakam 
| iha pancajhanasabdena kriyacaryayogayogottarayoga*niruttarani (em. : nirut- 
taras ca Cod.) tantrany ucyante. I have seen no occurrence in any Indian source 
of the term *Anuttarayoga, commonly encountered in secondary sources. It is ev- 
idently an incorrect modern translation into Sanskrit of the ambiguous Tibetan 
rendering of Yoganiruttara (rnal 'byor bla na med). Early authors attest a less 
developed hierarchy. Vilasavajra, an author of the eighth century (Tribe 1994, 
pp. 9-23) and the Guru of Buddhajnanapada according to Gzhon nu dpal (Blue 
Annals, p. 367), says that he writes his Namamantrarthavalokini after study- 
ing the Paramitanaya and the Kriya-, Carya-, and Yogatantras (A f. lvl-2: yo- 
gacaryakriyatantram tatha paramitanayam . . . vilokya), but the last evidently in- 
cludes texts such as the Guhyasamaja, Vajrabhairava, and Sarvabuddhasamayoga, 
since he quotes these and other related works. *Buddhaguhya (rNam par snang 
mdzad chen po mngon par byang chub pa'i rgyud chen po'i 'grel, ff. 64v7-65r6) 
speaks of Kriyatantras, which emphasize external ritual practice (phyi'i spyod, 
bahyacarya), giving as examples the Susiddhikara and the Vidyadharapitaka, and 
Yogatantras, which emphasize internal meditation (nanggi sbyor, adhyatmayogah), 
giving the example of the Sarvatathagatatattvasamgraha, and says that the Maha- 
vairocanabhisambodhi, later classified as a Caryatantra, is a Yogatantra in as much 
as it emphasizes the practice of Method and Wisdom (thabs dang shes rab gtsor 
gyur sbyor ba'i rgyud), but may also be referred to as a Kriyatantra or as an Ub- 
hayatantra (by a ba'i rgyud dam gnyis ka'i rgyud), that is to say, as a Tantra of both 
(ubhaya-) classes, because it also teaches external practice for the benefit of those 
whose commitment is to this. In a parallel treatment in his Pindartha commentary 
on the Mahavairocanabhisambodhi he gives the Vajrapanyabhiseka among exam- 
ples of Kriyatantras (see the translation in Hodge 2003, p. 449). This too was later 


The Saiva Age 

use of that tradition's system of the four types of Mudra (Mahamudra, Samaya- 
mudra, Dharmamudra, and Karmamudra) in Sadhana texts based on this 
Tantra, such as the Vajrajvalodaya of Anandagarbha and the Herukasddhana 
of Kalyanagarbha. 338 But it initiates a new direction that would be followed 
in the next and final phase of the Mantranaya's development, that of the 
Yoginitantras. 339 

considered to be a Caryatantra. The terms Kriyatantra and Yogatantra are seman- 
tically coherent, as Buddhaguhya indicates. But the choice of the term Caryatantra 
('Observance Tantra') for the intermediate class is puzzling. It is conceivable that it 
was adopted artificially under the influence of the classification of the subject mat- 
ter of the Tantras of the Saiva Mantramarga into kriyd, caryd, yogah, and jhdnam 
or vidyd, perhaps with the notion that the fourth corresponds to the Paramitanaya. 

1 As far as I am aware, only one other Sadhana text of this Heruka has sur- 
vived in Sanskrit. This is the anonymous Herukasddhana of Sddhanamdld 241. 
Anandagarbha's, which appears not to have been translated into Tibetan, is much 
the most detailed of the three. Apart from these works the only other evidence 
of this cult in surviving Sanskrit sources of which I am aware is in the eclectic 
Yogimtantra Samputodbhava, which in f. 80v5-81v2, in its eighth Kalpa, the Sar- 
vakriydsamudayakalpardja, includes the Mantras of this Heruka and his retinue 
of goddesses. There is also a chapter in the Abhidhdnottara of the Cakrasamvara 
corpus (B ff. 121v5-129vl: Patala 22) which teaches a hybrid pantheon in which the 
goddesses of this Heruka's retinue have been incorporated into that of Heruka and 
Vajravarahi, the former taking on the appearance of the Heruka of the Sarvabud- 
dhasamdyoga, being four-faced and eight-armed. This poverty of surviving sources 
in Sanskrit is probably due to the eclipse of this Tantra after the propagation of the 
later Yoginitantras, both in India and in Tibet. A striking indication of this eclipse 
is the fact that its Mandala was not included by Abhayakaragupta in his Vajrdvali 
and Nispannayogdvali in the first quarter of the eleventh century. For the position 
that the four Mudras are the distinctive fundamentals of the Sadhana system of the 
Yogatantras see, e.g., Mkhas Grub rje's rGyud spyi, pp. 228-248. 

' It was accordingly classified in the Kanjur (T5h. 366-367) among the Yoginitantras 
(Toh. 360-441). Likewise, Mkhas grub rje (1385-1438) in his rGyud spyi, p. 
266: bde mchog kye rdor dus 'khor sgyu thod gdan *bzhi (em. : gsum Ed.) 
phyag chen thig le sangs rgyas mnyam sbyor sogs ma rgyud yin no 'The Mother 
Tantras [=Yogimtantras] are such as the Samvara, the Hevajra, the Kdlacakra, 
the [Mahd]mdyd, the [Buddha]kapdla, the Catuspitha, the Mahdmudrdtilaka, and 
the [Sarva]buddhasamdyoga' '. This recognition of the [proto-]Yogimtantric char- 
acter of the text is not only Tibetan. It appears in the thirteenth chapter of the 
Ddkinivajrapanjara, where it is referred to in abbreviated form as the *Sarvabud- 
dha- (Sangs rgyas kun) in a list of Yoginitantras that also includes the Vajraddka, 
Hevajra, Guhyakosa, Vajrdmrta, and Cakrasamvara: rdo rje mkha' 'gro phan rgyud 
dang | *kye yi rdo rje (T : kye yi rdo rje dkyU 'khor D) sangs rgyas kun | gsang mdzod 
rdo rje bdud rtsi 'byung ba dang | 'khor lo sdompa gur *gyi (T : dang D) 'byung gnas 
ni | rnal 'byor ma *rgyud ni (T : rgyud drug tu D) rab tu grags (mKha' 'gro ma'i dra 
ba'i rdo rje gur rgyud, D f. 104v4-5; T p. 369, 11. 5-6), and in Dlpankarasrljiiana's 
commentary on his Byang chub lam gyi sgron ma'i dka' 'grel, where he refers to 
the texts of this class under their alternative title as Yoganiruttaratantras (rnal 
'byor bla na med pa'i rgyud), p. 286: rnal 'byor bla na medpa'i rgyud ni dpal nam 
mkha' dang mnyam pa 'bum pa chen po 'khor lo sdom pa dang rdo rje mkha' 'gro 
dang rdo rje gdan bzhi pa dang ma ha ma yd dang sangs rgyas mnyam sbyor 


Genesis and Development of Tantrism 

First, it introduces or brings to the fore the cult of the deity Heruka 340 with 
an iconography inspired by that of the Bhairavas of the Vidyapitha with their 
accoutrements and attributes of the cremation-ground dwelling Kapalika Saiva 
ascetic. According to the visualization given by Anandagarbha he has four faces 
and eight arms, emerging as the transformation of a dark blue flaming Vajra, it- 
self a transformation of a dark blue syllable HRIH. The central face is fierce (rau- 

dang sangs rgyas thod pa dang dgyes pa'i rdo rje bum phrag Inga pa la sogs pa 
rgyud sde stong phrag bcu gnyis bzhugs te rgyas par bya ba na grangs med do 
'The Yoganiruttaratantra, endless in its full extent, contains 12,000 [texts], princi- 
pally the Mahakhasama in 100,000 [verses], the Cakrasamvara, the Vajradaka, the 
VajracatuspTtha, the Mahamaya, the [Sarva]buddhasamayoga, the Buddhakapala, 
and the Hevajra in 500,000 verses'. On the term Yoganiruttara see here p. 146. 
340 The origin of the name Heruka has not been explained in a satisfactory manner. 
Indigenous sources explain it only through artificial semantic analyses based on su- 
perficial similarities of sound. Thus, for example, we are told that 'He-' means 'un- 
caused' (hetuvarjitam), '-ru-' means 'formless' (rupanirmuktam), and '-ka' means 
'free of sense-faculties' (karanojjhitam); see Vajrapani, Laghutantratika, p. 45; 
Bhavabhatta, Cakrasamvarapanjika, p. 5; and the Tibetans, who translated names 
if they were meaningful, either left this untranslated or substituted a description, 
namely Khrag 'thung 'Blood-drinker', a meaning that cannot be justified etymolog- 
ically. So if the name was meaningful at some stage it appears that that meaning 
has left no trace in the surviving literature. The alternative is that it never was 
meaningful in this sense, being created on the basis of the unmeaning syllables 
HE HE RU RU KAM that are found in Cakrasamvara's Mulamantra: OH SRIVAJRA 


may be said that the name appears without this doubling of the first two syllables in 
the earlier Sarvatathagatatattvasamgraha, section 794, in the Mantra for the tam- 
ing of all the Mother goddesses: OH HERUKA VAJRASAMAYA SARVADUSTASAMAYA- 
mudraprabhanjaka hum phat. It might seem more reasonable, then, to see he 
he ru RU kam as a spell-element built from an already existing name. However, it 
is striking that we find almost the same element in the Vidya of Parapara, an im- 
portant Mantra of the Sakta Saiva Vidyapitha: OM aghore hrih paramaghore 


RA PHAT HUM HAH PHAT (Siddhayogesvarimata 3.23-39; Malinivijayottara 3.42- 
50; Tantraloka 30.20-24b; Trisirobhairava quoted by Jayaratha thereon) and its 
variant taught in Kubjikamata 18.4-24: AIM AGHORE HRIM HSAH PARAMAGHORE 


he ru RU RA RA hrim hum phat. We may note that the name Hevajra, that of the 
second major deity of the Yogimtantras, appears to have a similar origin, having 
been conjured up from the Mantra he vajra pasya 'O Vajra[-being], behold!' that 
is uttered when the blindfold is removed from the candidate's eyes in the presence 
of the Mandala (Sarvatathagatatattvasamgraha, section 230). The origin of the 
Herukas Rigi-arali and Vajra-arali of the Tantras of those names are also, it seems, 
the apparently unmeaning syllables of Mantras: OM ARALI RIGI PHEM PHEM PHEM 
BHYO SVAHA (Ri gi a ra li'i rgyud f. 187v2) and OM VAJRA ARALI PHAT . . . PHEM 
PHEM SVAHA (Ri gi a ra li'i rgyud f. 187v7). The name of the Heruka Buddhakapala 
of the Tantra of that name has likewise been conjured out of the feminine vocative 
BUDDHAKAPALINI/-KAPALINI that appears in its Mantras; see (Nispannayogavali, 
p. 31: OM BUDDHAKAPALINI AH HI HAI HUM PHAT; Buddhakapalatantra, e.g., f. 
5rl: OM BUDDHAKAPALINI MATA 2 AH PHAT SVAHA puspanivedanamantrah). 


The Saiva Age 

dram), those to its right and left expressive of delusion and erotic passion, and 
that behind open-mouthed to devour. In his two uppermost hands he holds the 
freshly flayed skin of Bhairava over his back, in the two below a bow and arrows, 
in the third right in descent he shakes a blazing three-pronged Vajra, and in 
the fourth a skull-bowl filled with human blood (mahdraktam). In the third left 
in descent he brandishes the Kapalika's skull-staff (khatvdngah), topped with a 
three-pronged Vajra and adorned with bells, and in the fourth a skull-bowl filled 
with human flesh (mahdmdmsam). Or he may be single-faced and two-armed, 
with a five-pronged Vajra in his right hand raised above his shoulder and a skull- 
bowl full of human flesh in his left, with a skull-staff resting on his left shoulder 
and held in the crook of his left arm. He wears a chaplet of skulls with the Bud- 
dha [Aksobhya] adorning his flaming hair, is surrounded by an aureole of flames, 
poses with his left foot on the ground and his right leg raised so that the sole of 
the foot touches his left thigh, has dancing eye-brows knitted in anger, and has 
round, fire-red darting eyes. 341 Kalyanagarbha, who teaches only the two-armed 
form, adds that he stands on a sun disc, which rests on a lotus, which rests in 
turn on a prostrate corpse, is smeared with ashes, wears a garland of freshly sev- 
ered human heads, and has protruding fangs. 342 An anonymous Sadhana text, 

341 Vajrajvdlodayd, f. 172vl— 2: bhagavato mahdmudrdm baddhvd purata dkdsadese 
HRl<B.>kdrena visvapadmam nispddya tasyopari pancasucikam jvdldvajram HUM 
A iti | tato vajrdhamkdra<m> bhdvayet JVALAVAJRO 'HAM HUM iti \ tatas tad 
vajram sriherukam dtmdnam bhdvayet SRlHERUKO 'ham HUM iti; f. 173r4-v4: 
caturmukham astabhujam | tatra prathamam mukham raudram daksina<m> 
dvitiya<m> mukham pramoha*pramodina<m> (?) prsthatas trtiyakam bhaksana- 
mukham vdmatas caturtham srhgdramukham | etac ca mukha*catustayam (conj. 
Isaacson : catustaya Cod.) gltyd nirdistam iti | dvdbhydm bhujdbhydm vdyu- 
patadhdranayogena sdrdrabhairavacarmadharam dvdbhydm dhanurbdnadharam 
daksinatrtiyena trisucikajvdldvajrolldlanatatparam caturthena mahdraktapari- 
purnakapdladharam vdmatrtiye ghantdsahitavajrakhatvdhgadharam caturthena 
mahdmdmsaparipurnakapdla :, 'dharam (corr. : dharah Cod.) | dvibhujam eka- 
mukham <vd> vdmaskandhe yajhopavitayogena ghantdvajrakhatvdhgasobhitam 
daksinakarena *tripatdkdyuktena (corr. : tripatdka Cod.) pancasucijvdld- 
vajradharam | vdmakarena mahdmdmsaparipurnakapdladharam | kapdlamdld- 
makutabuddhacuddmani<m> uccavisvapadmdsanopavistam vdmapddam bhumi- 
stham krtvd daksinapdda<m> sattvaparyahkayogena nyasya | tatpddatalam 
vdmorund samputlkaranayogendvasthdpya nllajvaldvajramayam raktajvdldbha- 
mandalam mahdpralayakdlograsmasdndgnisadrsam diptakesam raudrddirasa- 
samyogavicitramukhavibhramam | savibhramabhrubhrkuti<m> pradiptdloka- 
nartitadrstim iti. 

342 Kalyanagarbha, Herukasddhana , pp. 470-471: adhomukhasya savasyopari visva- 
padmam tasyopari suryamandalam tanmadhye samupavistam *ekdsyordhvabhuja- 
dvayam (ekdsyo em. : ekasyo Ed.) iti vacandd ardhaparyahkinam bhasmoddhu- 
lita<m> raktaprabhdmdlinam pingalordhvakesam . . . sdrdranaramastakamdld- 
krtasragddmam damstrdkardlavadanam caladvartuldkdraraktdksam savibhrama- 


Genesis and Development of Tantrism 

which also teaches only that form, gives the further details that he is dark blue 
and clad in a garment of human skin, that his garland of heads is strung together 
with human entrails, that he is adorned with human bones, that is to say with 
the Kapalika ornaments known as the Mudras, and that his posture indicates 
that he is dancing. 343 

He is surrounded in the style of the Vidyapitha by twenty Vajradakinis: 344 
first, in the innermost circuit the eight Gauri, Cauri, Pramoha, Vetali, PukkasI, 
Candali, Ghasmari, and Herukasamnivesa/Herukasamnibha; then the four 
Capadharini, Khatvarigadharini, Cakradharini, and Citrapatakadharini; then 
four offering goddesses: Puspa, Dhupa, Aloka, and Gandha; and finally four 
theriocephalic gate-guardians: Turarigama, Vajramukhi, Vajramamaki, and 
Bhasmapralayavetali. 345 

343 Sddhanamdld no. 241: tato hrihkdranispannam nilakardlavajram hrihkdrddhi- 
sthitavaratake dhydtvd tatsarvaparinatam nilam naracarmabhrtam kapdlamdld- 
ksobhyasiraskamjvaladurdhvapihgalakesam raktavartuldksamantrasamgrathita- 
mundamdldvalambitam nardsthiracitdbharanam dvibhujaikamukham damstrd- 
kardlavadanam . . . visvapadmasurye vdmapddam tasyaivorau daksinacaranam 
vinyasya nrtyam kurvantam herukaviram bhdvayet. There are numerous two- 
armed Herukas conforming to the iconographical prescriptions of these Sadhanas 
in surviving statuary from eastern India, though this connection with the tradition 
of the Sarvabuddhasamayoga has not been recognized to my knowledge. For ex- 
amples from Ratnagiri in Orissa, Nalanda, Sarnath, and Subhapur (in the Comilla 
District of Bengal) see Linrothe 1999, pp. 249-260, figs. 175-183, and 185-188, 
and Huntington 1984, fig. 215. The last lacks the prostrate corpse. 

344 Vajrajvalodaya, f. 176r7-vl: sarvam srlgauryddivajradakiniganam nirmdya 
prajvalitordhvakesam | raktajvdldbhamandalam mahdpralayakdlograsmasdndgni- 
sadrsam samkruddham ekakapdlaikabuddhamakutam svacihnadharam yathd- 
sthdne nivesayet. 

345 The Sarvabuddhasamayoga deploys a complex six-family Mandala consisting of six 
sub-Mandalas. The six families, each with its own sub-Mandala, are those Va- 
jrasattva, Vairocana, Heruka, Padmanartesvara, Vajrasurya, and Paramasva. Two 
Mandala traditions deploy this pantheon. In one Vajrasattva occupies the cen- 
tral sub-Mandala and in the other Heruka. In each sub-Mandala one of these 
six occupies the centre surrounded by twenty goddesses. The last twelve god- 
desses are the same in each, namely Susira, Nrtya/Vina, Vitata, and Ghana, 
followed by Puspa, Dhupa, Aloka, Gandha, Turaga, Vajramukhi, Vajramamaki, 
and Bhasmapralayavetali, the first eight of these being, as their names reveal, 
offering-goddesses (pujddevyah), personifications of offerings, and the last four gate- 
guardians, except that in the retinue of Heruka Capadharini, Khatvangadharini, 
Cakradharini, and Citrapatakadharini are substituted for the first four, the mu- 
sical offering-goddesses Susira, Nrtya/Vina, Vitata, and Ghana. The first eight 
of the twenty, then, stand apart as the retinue specific to each Tathagata. The 
eight from Gauri to Herukasamnivesa formed the basis of the retinue of Hevajra 
in the Yogimtantra Hevajra, with the difference that there we see Sabari rather 
than Pramoha and Dombi rather than Herukasamnivesa. See TOMABECHI 2007, 
pp. 919-921 for a complete tabulation of all one hundred and twenty-six deities and 
their seed-syllables as given in the Sarvabuddhasamayoga and the Paramddya. 


The Saiva Age 
According to Anandagarbha 346 Gauri (E) is fair in colour and tranquil-faced. 

See also Tanaka 1996, pp. 199-201 for the Tibetan names of all the goddesses (and 
their Mantras) in the six sub-Mandalas, and the listings of the names and posi- 
tions of all the deities of the two six-family Mandalas in Bsod NAMS RGYA MTSHO 
1991, pp. 106-113. In the Heruka-centred Mandala set out there each of the six 
deities presiding over the sub-Mandalas has a consort: Heruka + Isvari, Vairocana 
+ Locana, Vajrasurya + Mamaki, Padmanartesvara + Pandaravasim, Paramasva 
+ Tara, and Vajradhara + Samvari; and the total of deities is 135, since two extra 
goddesses, counted as one, Citrapadma and Citravajra, are found in front of the cen- 
tral deity in the sub-Mandala of Paramasva, and there are eight additional deities 
in the outer enclosure, since there too there are four offering goddesses within its 
corners and four animal-headed goddesses guarding its gateways. Theriocephalic 
female gate-guardians are a common feature in the Mandalas of the Yogimtantras; 
see, e.g., Samvarodaya 13.29c-31b; Jayabhadra, Cakrasamvarapahjikd, p. 113 on 
2.8 (Kakasya, Ulukasya, Svanasya, Sukarasya); Nispannayogdvali, p. 15 (Hayasya, 
Sukarasya, Svanasya, and Simhasya in the 17-deity Mandala of Hevajra) and p. 
90 (Sukarasya, Grdhrasya, Jambukasya, Garudasya, Vyaghrasya, Ulukasya in the 
Mandala of Kalacakra). 
3 Vajrajvdlodayd, ff 177r4-178r5: purvadigbhdge gauri gauravarnd sdntadrstih 
saumyamukhd yaugapadyenaiva tiksnadhanurbdnapariksepdn mahdprasahya- 
sira<s>catustayam patayanti pratydlidhasthdnasthd \ daksine cauri raktavarna 
raudradrstimukhd yajnopavitayogena vdmaskandhe khatvdhgam dhdrayanti \ 
kapdlamalamukuta vdmakrodhamustina hrdy ahkusadharini daksinakarena 
madhyahgulyastaracakram utkarsayanti vdmapddena trailokyam lahghayanti \ 
pastime pramoha adivarahamukhd pramohadrstih krsnd caturbhuja madya- 
purnakapdlavdmakara daksinakare vajradhdrini punar vdmadaksinabhujdbhyam 
*parasparabaddhabhydm (corr. : paramparabaddhabhyam Cod.) prthivy- 
uddharanam kurvanty alidhapaddvasthitd \ uttare vetali sitavarnam harsa- 
mukhim mrtakotthapanadrstih daksinakarena candrakantabhakapalacasakena- 
mrtavaridharam patayantim vamakarena vajrapatakakaradharinim yathesta- 
padavasthita | tasminn eva mandale purva*kosthe (corr. : kostha Cod.) pukkasi 
visvavarna nrtyamukhl nrtyadrstih daksinavajramustina pancasucikajvalavajra- 
dharini \ vamakarena marutoddhutakalpavrksalatadharini kapalamaladipari- 
purnasadhumasmasanamadhye nrtyaprayogena | daksine candall nilavarna vata- 
mandalikarudha savibhramamukhi urdhvadrstih daksinamustina vajrasulam 
dddya \ vayupatadharanena vatamandalikapramoksena sddhya*prandmadayo 
patanti (?) | pastime ghasmari krsna*varna (corr. varnnam Cod.) mrta- 
carvanamukhi bhaksanadrstih \ vamakarena vajrajvalagnikundadharini \ daksine 
vajramustina khadgam dddya pratydlidhapaddvasthitd | uttare sriherukarupa- 
samnibha vamakarena *casakakapdlam (casaka conj. : capdsa Cod.) dddya 
vdmaskandhe khatvdhgam dhdrayanti | daksine tripatdkdkarena pahcasucika- 
jvdldvajram dddya srlherukapade dvibhujaikamukhl samsthitd | dgneyakosthake 
*capadharini (em. : copodhdrini Cod.) | raktavarna vamakarena vajra- 
dhanur dddya daksinena *vajracdpasahitena (corr. : vajracdpdsahitena Cod.) 
dhanu<r>gundkarsanayogena *vajrabdndn (corr. : vajrabdrnndn Cod.) ksipanti | 
nairrte khatvangadharini kapdlamdldmakutabuddhacuddmani<r> *drsitdra (?) 
bhasmasubhravarnd daksinakarena ca pahcasutikajvdldvajra<m> pdnyd ksipanti 
| *vdyavye (em. : vdyave krodhamustind tarjanitatpard | vdyavye Cod.) *cakra- 
dharini (corr. : cakradhdri Cod.) gauraharitavarnd vdmakrodhamustina tar- 
janatatpard *daksinakaramadhyamdhgulydstdracakram (daksina corr. : daksine 
Cod.) utkarsayanti | aisdne kone citrapatakadharini \ *kanakopalavarnd (varnd 


Genesis and Development of Tantrism 

Eight-armed, she cuts off each of the four heads of Brahma by simultaneously fir- 
ing arrows from four bows. 347 Cauri (S) is red and fierce-faced. Wearing a chap- 
let of skulls she holds a goad-hook (ankusah) in her left hand at her heart with 
a skull-staff in the crook of her left arm resting on her left shoulder, and holds 
aloft an eight-spoked discuss with the middle finger of her right, pressing down 
on the three worlds with her left foot. Pramoha (W) is black and four-armed, with 
the face of Visnu's boar-incarnation (ddivardhamukhd). In her first left hand she 
holds a skull -bowl full of wine and in her first right a Vajra. With her other two 
hands she imitates the boar-incarnation by raising up the earth. 348 Vetali (N) 
is white and joyful-faced. With her right hand she pours a stream of the nec- 
tar of immortality from a transparent skull-cup and with her left shows the Va- 
jra banner gesture. Pukkasi [E] is multi-coloured (visvavarna) and dancing in a 
smoky cremation -ground full of strings of skulls and the like. In her right fist she 
clasps a five-pronged Vajra and in her left a wind-buffetted tendril from the wish- 
granting tree of paradise (kalpavrksalata). Candali (S) is dark blue and riding 
on a whirlwind (vatamandalika). In her right fist she clenches a Vajra-topped 
trident and with her left releases a whirlwind against her victims. Ghasmari 
(W) is black and eating a corpse. In her left hand she holds a blazing sacrificial 
fire-vessel (agnikunda-) and with her right grasps a sword. Herukasamnibha 
(N), black like Heruka, holds a skull-cup [to her heart] in her left hand, with 
a skull-staff resting on her left shoulder, and a five-pronged Vajra in her right. 
Capadharini (SE) is red and, holding a Vajra bow with her left hand, fires Vajra 
arrows by drawing back the bowstring with her right. Khatvarigadharini (SW) 
is ash-white, wearing a chaplet of skulls and the Buddha on her crown, [holding 
a skull-staff with her left hand and] hurling a blazing fire-pronged Vajra from 

conj. : varnna Cod.) daksinakarena *samghata(l)vicitravarnapataka<m> dhara- 

347 That Gauri is eight-armed is not stated by Anandagarbha, but she could not draw 
four bows simultaneously with fewer and no other hands are mentioned. His 
mahdprasahya- is obscure but evidently it denotes Brahma since the victim is 
said here to have four heads (mahaprasahyasira<s>catustayam patayanti). Both 
these inferences are supported by Humkaravajra, who is explicit in both regards 
in his *Herukasadhana (f. 203v2): zhal bzhi phyag brgyad brjid pa'i stongs \g.yon 
brkyang gar gyis bzhugs mdzad cing \ mda' bzhi dus gcig bkang ba la | tshangs pa'i 
tngo bzhi spyangs pa ste. 

348 According to Humkaravajra's Herukasadhana she has two heads, that of a boar 
above and a red head below. Moreover, he has her raise with her two lower hands 
a wheel ('khor lo) rather than the earth (f. 203v3-5): *pra (em. : bra Cod.) mo dbu 
gnyis gong ma phag | 'og ma dmar po phyag bzhi pa | g.yas kyi dang pos rdo rje 
rtse gsum bsnams | g.yon gyi dang pos kham phor 'chang | 'og gnyis khu tshur so sor 
'chang | 'khor lo 'dzin cing bteg pa'i tshul | g.yas brkyang stabs bcas nub phyogs su | 
rmongs tshul mdog dmar pa dma la. 


The Saiva Age 

her right. Cakradharini (NW) is light green and holds aloft an eight-spoked dis- 
cuss on the middle finger of her right hand and threatens [the wicked] with her 
left fist clenched in anger. Citrapatakadharini (NE) is golden in colour, holding a 
multi-coloured banner in her right hand. The four offering-goddesses stand in the 
directions holding the offerings that they personify: flowers, an incense-burner, 
a lamp, and fragrant powder; and the four goddesses Turangama, Vajramukhi, 
Vajramamakl/Aloka, and Bhasmapralayavetali stand in the four gates of the en- 
closure to subjugate all hostile deities (krodhakulam), with the heads of a horse, 
a boar, a crow, and a dog, and holding a hook, noose, chain, and bell. 349 

All this, barring a few specifically Buddhist details such as the Vajras 
and the offering-goddesses, who are already in the Mantranaya of the Sarva- 


Anandagarbha's text is corrupt and lacunose at this point in the manuscript, 
omitting Vajramukhi and Vajramamaki (f. 178r5-v2): vamamustina Ighatylava- 
sthita Itrylamanjalina puspadhupadlpagandhacihnadharinyah asvagojasabhuti- 
samjnitasattapujadevi | purvadvaramadhye turangasana vamahastena padma- 
hastd hayagrlvaharitam asvamukham dharayanti | daksine kare sthitena vajra- 
hkusena sarvakrodhakulam akarsayantl | pastime *dvara alokam (corr. : dvare 
lokan Cod.) candrasuryamandala?ru?payuktavajra*sphotanena (conj. : sphotanam 
Cod.) sarvam krodhakulam bandhayanti \ uttaradvare bhasmapralayavetali 
*vamakarena (corr. : namahkarena Cod.) kapalamadhye visvavajrastham buddha- 
bimbam dharayet | daksine kare sthitavajraghantavadanayogena sarvakrodha- 
kulam vasikurvanty *avasthita (corr. : avasthitah Cod.) | *sarvas caitah (corr. 
: sarvvancetah Cod.) pratyalidhasthanastha<h> sadrstibhavarasanvita<h>. A 
complete but less detailed description of these eight can be seen in the Tibetan 
translation of the *Herukasadhana of Humkaravajra, f. 204r4-7. The identity of 
the non-human heads of the gate-guardians is mentioned in these sources only in 
the case of the horse-headed Turangama, by Anandagarbha and Humkaravajra (f. 
204r5: shar sgo rta mgrin 'phang mtho dkar | g.yas na rta gdong g.yon Icags kyu), 
and Vajramukhi, by Humkaravajra, who names this goddess Phag gdong 'Boar- 
face' (Sukarasya) (f. 204r6: Ihor sgor phag gdong snon mo ste \ g.yas pas mche ba 
g.yon zhags 'dziri). According to the tradition of the Ngor Mandalas, the last two 
door-guardians, Snang ba ma (Aloka) and Thai byed ma (*BhasmakarinI [?]), are 
crow-faced and dog-faced (Bsod nams rgya mtsho 1991, p. 110). These animal- 
headed guardians exemplify the character of this Tantra as transitional between 
the Yogatantras and the Yogimtantras. The animal-headedness is shared with 
such goddesses in the latter (see here p. 151), but the hand-attributes, namely the 
hook, noose, chain, and bell, are those of Vajrankusa, Vajrapasa, Vajrasphota, and 
Vajravesa, the male gate-guardians of the Vajradhatumandala of the Yogatantra 
Sarvatathagatatattvasamgraha; see Tanaka 1996, p. 271. For those attributes see 
the *Herukasadhana of Humkaravajra, f. 204r5-7 (I have restored the Mantras, 
which invoke the goddesses as the personifications of these attributes, to their cor- 
rect Sanskrit form): OM *vajrankuse (corr. ba dzra am ku sha Cod.) JAH | shar 
sgo rta mgrin 'phang mtho dkar | g.yas na rta gdong g.yon Icags kyu | OM VAJRAPASE 
HUM | Ihor sgor phag gdong sngon mo ste | g.yas pas mche ba g.yon zhags 'dzin | OM 
*VAJRASRNKHALE (corr. : BA DZRA SHRI KHA LE Cod.) VAM | nub sgor snang byed 
dmar mo ni \ phyag gnyis nyi zla Icags sgrog 'dzin \ OM *VAJRAGHANTE (corr. : BA 
DZRA GA NTE Cod.) HOH | byang sgor thai byed mdog Ijang du \ sang rgyas gzugs 
dang dril bu'o. 


Genesis and Development of Tantrism 

tathdgattattvasamgraha, 350 is very much in the Kapalika style of the pantheons 
of Bhairavas and Yoginis taught in the Vidyapitha. 

Second, it is in the tradition of this Tantra that we see for the first time 
in the Mantranaya the practice of the ganamandalam, orgiastic worship in an 
assembly consisting of a male and a group of female adepts (yoginiganah) person- 
ifying the deities of the cult, with a jargon of special terms and gestures known 
as chommah to be used in these gatherings. 351 Both these features, collective 
orgiastic worship of deity-personifying Yoginis and the use of chommah, are dis- 
tinctive features of the Sakta Saivism of the Vidyapitha. 352 

Third, we see here for the first time the complete abandoning of the mixed 
prose and verse style inherited from the Mahayanasutras in favour of one that 
resembles that of the Saiva scriptures in consisting entirely of Anustubh verse, 
barring the Mantras, and also the disappearance of the traditional Buddhist 
preamble maintained up to the time of the Guhyasamdja, stating the occasion 
and place of the revelation. 353 It is also in the Sarvakalpasamuccaya, the supple- 

350 See Sarvatathagatatattvasamgraha , sections 165-176 and Nispannayogavali, p. 46 
(Vajradhupa, Vajrapuspa, Vajraloka, and Vajragandha). 

351 The practice and the jargon are outlined by Aryadeva in his Caryameldpakapradipa 
(pp. 82-60: prapancatdcaryd) on the authority of this Tantra. The Yoginis per- 
sonified here are the twenty that form the retinue of Vajrasattva, the eight pecu- 
liar to him being Samvari, Ahosukha, Pradipa, Sisya, Buddhabodhi, Dharmacakra, 
Trailokya, and Kamalata. 

352 On such worship in Sakta Saivism see Sanderson 2007a, pp. 280-288; and 
Tantrdloka 28.6-111, 372c-385b (yoginimelakah, cakraydgah, murtiydgah), 29.66, 
78-79. On chommah in these traditions see Sanderson 2007a, p. 333 and the 
sources quoted in footnotes 331-332. 

353 The Tantra begins as follows (Sangs rgyas thams cad mnyam par sbyor ba, 
f. 151rl-2: 1.1 sems dpa' sangs rgyas kun gyi dngos | rdo rje sems dpa' bde 
ba'i mchog | gsang ba mchog gi dgyes pa na | thams cad bdag nyid rtag tu 
gzhugs | 1.2 'di ni rang byung bcos Idan 'das \ gcig bu rob tu phye ba'i lha \ 
sangs rgyas thams cad mnyam sbyor ba | mkha' 'gro sgyu ma bde ba'i mchog 
(*rahasye parame ramye sarvatmani sada sthitah | sarvabuddhamayah sattvo 
vajrasattvah param sukham || asau svayambhur bhagavan eka evddhidaivatah || 
sarvabuddhasamayogadakinijalasamvarah). Cf. the opening verses of the Laghu- 
samvaratantra, which are evidently based on it: athdto rahasyam vaksye samasan 
na tu vistarat | sriherukasamyogam sarvakamarthasadhakam || 1.2 uttarad api 
cottaram dakinijalasamvaram | rahasye parame ramye sarvatmani sada sthitah 
|| 1.3 sarvadakinimayah sattvo vajrasattvah param sukham | asau hi svayambhur 
bhagavan viro dakinijalasamvaram; and the following citation of the Sarvabuddha- 
samayoga in the Caryamelapakapradipa, p. 82: athatah sampravaksyami sarvato 
visvam uttamam | sarvabuddhasamayogam dakinijalasamvaram || rahasye parame 
ramye sarvatmani sada sthitah | sarvabuddhamayah sriman vajrasattvodayah 
sukhah. These verses are 1-2 of the Kalpa 6 of the Tantra, corresponding to 
the Tibetan, except that that seems to have had a different version of the first 
line (f. 159v4-5): de nas gzhan yang thams cad du | rnam pa sna tshogs 
mchog 'byung pa'i \ sangs rgyas thams cad mnyam sbyor ba \ mkha' 'gro sgyu ma'i 


The Saiva Age 

mentary continuation {uttaratantra) of this Tantra, that we see the first appear- 
ance in the Mantranaya of the Saiva method of teaching Mantras in encrypted 
form to be decoded by the process known as mantroddharah; and with this devel- 
opment we encounter what is at present our earliest evidence of Buddhist-Saiva 
intertextuality A passage of seven verses that prescribes for this purpose the 
drawing of a square with forty-nine cells (kosthakani) and the arranging of the 
forty-nine letters within them corresponds very closely to one in the Vlndsikha 
of the vdmasrotah division of the Vidyapitha. 354 

The intensification of the Sakta Saiva character of the Mantranaya evident 
in this text is accompanied by the implication that this Buddhism is one that 
has conquered that tradition, transforming it, as it were, from within into a ve- 
hicle for Buddhist salvation. For while wrathful Heruka appears with Kapalika 
iconography and a retinue of Yoginis he wears, as we have seen, the freshly flayed 
skin of Bhairava over his shoulders; and the Tantra relates that its deity in its 
commitment to purify all beings has violently overpowered Siva, Visnu, Brahma, 
and Kamadeva, and taken their consorts by force for his own enjoyment. 355 This 

bde mchog bshad. For the requirement of a preamble see, e.g., the Mahayanist 
Dharmasamgitisutra as quoted by Abhayakaragupta in the introduction to his Ab- 
hayapaddhati f. lv: kdladesadesakaparsatsdmagri hi desandyd niddnam endm 
vind desandnupapatteh. tatra evam mayeti mama dharmah samgatavyah . . .ity 
uktam bhagavatd dharmasamgitisutre 'For the preamble that establishes the au- 
thenticity of a teaching [comprises] all these factors together, namely the time, 
place, teacher, and congregation, because without all those it cannot be [accepted 
as] a teaching. To this effect the Buddha has declared in the Dharmasamgitisutra 
. ..: 'My teachings must be recited with [the opening phrase] "Thus I [. . . ]"'; and the 
unnamed Sutra quoted by Tathagataraksita on Yoginisamcdra 1.1: mayi parinirvrte 
bhiksava evam mayetyddikayd mama dharmah samgatavyah 'O monks, after I have 
been completely extinguished [by death] you should recite my teachings with the 
words "Thus I . . . "'. 

354 This nas b een demonstrated in Tomabechi 2007. The Saiva passage is Vlndsikha 
52-58. That in the Sarvakalpasamuccaya is DK, Rgyud 'bum, vol. ka, ff. 194v6- 

355 Samvaratantra (= Sarvabuddhasamdyogaddkinijdlasamvara) quoted in 
Jndnasiddhi 18.10—18 (pp. 153—154): sarvasuddhyadhimoksena prasahya 
balavdn adhah | pardkramakramandt tu sarvalokdn pramardayet || anyam 
tu dustaraudrogram sattvadhdtum anekadhd || pdpais corair avaskandhaih 
sarvam eva visodhayet | cchalena mdyayd caiva prasahya balavdn adhah || 
pancdyudhanibandhais ca sarvalokdn jayet tadd | vijitya sakaldm siddhim jagat 
sthdvarajahgamam \\ vicitravinayopdyaih svapardn anupdlayet | kdminindm 
bhavet kdmo raudrdndm raudram uttamam || saumydndm paramam saumyam 
hathdndm hathavikramah | paramesarn samdkramya prasahya balavdn adhah 
|| umadevim samdkrsya copabhogair bhunakty asau \ narayanam samdkramya 
prasahya balavdn adhah \\ rupinim. tu samdkrsya upabhogair bhunakty asau 
| prajapatim. samdkramya prasahya balavdn adhah \ prasantadevim dsddya 
upabhogair bhunakty asau \\ kamadevam. samdkramya prasahya balavdn adhah 
| ratiprltidhrtyaisvaryam samdkramya bhunakty asau. This corresponds to 


Genesis and Development of Tantrism 

rhetoric of appropriation is reflected in the Mantras of Heruka's Vajradakinls. 
Pramoha, who, as we have seen, has the boar face of Visnu's Adivaraha incarna- 
tion, is invoked as Vajranarayani, Cauri as Vajracandesvari, and Ghasmari as 
Vajramahesvari. 356 Furthermore, Heruka's first appearance in the Mantranaya 
is in the Sarvatathagatatattvasamgraha, where his name appears in a Mantra 
for the drawing of all the [Saiva] Mother-goddesses into Buddhism, and it is 
that, with the insertion of a single seed syllable, that is adopted as the Mantra 
of Heruka in the Sarvabuddhasamayogadakinijalasamvara. 351 The very title of 
the work alludes to this assimilation, since it is evidently calqued on those of two 
Vidyapitha scriptures, the Sarvavlrasamayoga and the Yoginijalasamvara. 358 

The Yoginitantras and the Full Appropriation of Vidyapitha Saivism 

With the Yoginitantras proper we reach the final stage of this process of 
absorption. The principal among the numerous Tantras of this class are the 

Sangs rgyas thams cad mnyam par sbyor ba, ff. 158v7-159r5, except that there 
Paramenia's (Siva's) consort is Bhimadevi (f. 159r2: lha mo bhi mo) and Narayana's 
(Visnu's) is Rukmini (f. 159r3: ru gmi ni). 

356 Vajrajvalodayd, f. 176v: HUM VAJRANARAYANI JHIR iti (em. : jhirati 

kapalamalamukute rulu rulu hum iti caurim. Ghasmari is invoked as Va- 
jramahesvari in the Mantras of the retinue of Heruka given in the Samputodbhava 


marydh (f. 81r4-5). There are other examples of the assimilative transformation 
of non-Buddhist deities in the Mantranaya, marked, as here, by the prefixing of 
Vajra- to their names. For example, the deities Vajranarayana, [Vajra]candisvara, 
and Vajrapadmodbhava, that is to say, Vajrayanist transformations of Visnu, Rudra, 
and Brahma, together with their consorts Vajrasri, Vajragauri, and Vajratara, join 
Akasagarbha and Khavajrini to form the retinue of Vajrasattva in the central sec- 
tion of the abridged Mandala (bsdus pa'i dkyil 'khor) of the Yogatantra Paramadya, 
a text with which the Sarvabuddhasamayoga is closely related (TOMABECHI 2007, 
p. 904; Tanaka 1996, pp. 271-272). That disposition of deities is taught (see 
Tanaka 1996, pp. 96-103) in the mChog dang po'i sngags kyi Hog pa'i dum 
bu (*Sriparamadyamantrakalpakhanda) (T5h. 488) according to Anandagarbha's 
mChog dang po'i rgya cher bshad pa (*Paramadyatika) (Toh. 2512). 

357 Sarvatathagatatattvasamgraha, section 794: OM HERUKA VAJRASAMAYA SARVA- 
dustasamayamudraprabhanjaka hum phat sarvamatrnam iti; Vajra- 
jvalodayd: OM herukavajrasamaya h<r>Ih sarvadustasamayamudra- 
PRABHANJAKA HUM PHAT iti svamantrena sriherukam nivesayet. 

358 On these two scriptures see Sanderson 2007a, pp. 234-236 and footnotes 21- 
22. The expression sarvavTrasamdyogaddkinTjdlasamvaram, without the substi- 
tution of -buddha- for -vira-, is seen in the Yoginitantras of Cakrasamvara. It ap- 
pears in, e.g., Laghusamvaratantra, f. 8r3 (8.1) and f. 24v4 (31.13ef): tatah sarva- 
virasamdyogaddkinljdlasamvaram; and Samvarodaya 3.6cd: sarvavirasamdyoga- 
ddkinijdlasatsukham . In the last satsukham is a tacit semantic analysis of 


The Saiva Age 

Laghusamvara also called Cakrasamvara and Herukdbhidhdna, the Heva- 
jra, the Catuspltha, the Vajramrta, the Buddhakapdla, the Mahdmdyd, the 
Rigydrali, the Vajrarali, the Candamahdrosana, and the Kdlacakra. Two of 
these texts, the Laghusamvara of the Heruka called Samvara (bDe mchog) 
or Cakrasamvara ('Khor lo sdom pa) and the Hevajra of the Heruka Hevajra 
held centre-stage, a position they later shared with the Kdlacakra when that 
text was propagated towards the end of the tenth century, during the reign 
of Mahipala I (r. c. 977-1027). 359 Their importance is reflected in the shere 
quantity of commentaries and other texts devoted to the cult of their deities. 
The Tenjur contains translations of eleven commentaries on the Hevajra and of 
eleven on the Laghusamvara, and of about two hundred other explanatory texts 
related to each. Moreover, they both have a number of satellite Tantras, the 
Hevajra five and the Laghusamvara over fifty 360 The principal among these, 
those that received commentaries, are for the Hevajra the Ddkinlvajrapanjara 
and the Mahdmudratilaka, and for the Laghusamvara the Herukdbhyudaya, 
the Vajraddka, the Abhidhdnottara, the Yoginlsamcdra, the Samvarodaya, and 
the Ddkdrnava. Another major Yoginitantra, the Samputodbhava, on which 
we have an important commentary, the Amndyamahjarl, by Abhayakaragupta 
( 1064-1 125), 361 pertains to both cycles. 362 

359 On the date of the Kalacakra see here p. 96. On the establishing of this tradition 
and how it positioned itself in relation to earlier Tantric Buddhism see Sferra 

360 This i ar g e total includes thirty-four texts (Toh. 383-416), forming a supplementary 
collection, as it were, of related opera minora, totalling less than 150 pages. Though 
included in the Kanjur they were classified by Bu ston (1290-1364) as supplemen- 
tary Tantras whose authenticity, that is to say, Indian origin, was the subject of 
debate (rgyud yang dag yin min rtsod pa can). The great majority are claimed in 
their colophons to be translations prepared in the early eleventh century by 'Brog 
mi in collaboration with the Indian Gayadhara. On the lay Tantric Gayadhara, who 
is mentioned in no Indian source known to me but is the subject of many partly 
conflicting accounts in Tibet, where he was venerated as the Indian source of the 
Lam 'bras tradition and for having collaborated with several Tibetan translators, 
see Stearns 2001, pp. 47-55. It is, however, certain that not all these opera mi- 
nora are of suspect authenticity. For my pupil Peter-Daniel Szanto has recently 
identified the original Sanskrit of one, the Anavilatantra, among the contents of a 
palm-leaf codex preserved in the Tokyo University Library (verbal communication). 

361 These dates rest on Tibetan tradition and are consistent with the regnal years of 
Ramapala that Abhayakaragupta has reported as the dates of composition at the 
end of some of his works; see here p. 126. 

362 Thus, though counted as an explanatory Tantra of the Cakrasamvara cycle, it is 
grouped with the Hevajra and Ddkinlvajrapanjara as one of the three Tantras of 
Hevajra (kye rdo rje rgyud gsum) in the Sa skya tradition of Tibet, and classified 
because of its mixed character as the Hevajra's shared explanatory Tantra (thun 
mong bshad rgyud); see STEARNS 2001, pp. 173-174, n. 28. It also extends into 
the territories of the Catuspltha, the Guhyasamdja, the Vajrabhairava, and, as we 


Genesis and Development of Tantrism 

Chronology and Provenance. All of these Tantras were translated into 
Tibetan, and all but the latest among them, the Dakarnava and the Samvarodya, 
were translated in the first half of the eleventh century, during the opening phase 
of the later diffusion (phyi dar) of Indian Buddhism to Tibet, as were commen- 
taries on the majority of those named here, most written during the course of the 
tenth and early eleventh centuries. 

The oldest is probably the commentary of Jayabhadra on the Laghusamvara. 
In chapter 38 of his Rgya gar chos 'byung Taranatha includes five of our com- 
mentators on the Laghusamvara, Jayabhadra, Bhavabhadra/Bhavabhatta, 
Bhavyakirti, Durjayacandra, and Tathagataraksita, among ten persons whom 
he holds to have occupied the office of chief Vajracarya at Vikramasila in rapid 
unbroken succession, and claims that Jayabhadra was the first of the ten 
(Jayabhadra, Sridhara, Bhavabhadra (/Bhavabhatta), Bhavyakirti, Lilavajra, 
Durjayacandra, Krsnasamayavajra, Tathagataraksita, Bodhibhadra, and 
Kamalaraksita). Moreover, comparison of the commentaries, the Tibetan trans- 
lation, and the only manuscript of the Laghusamvara accessible to me at present 
reveals two versions of the text. Taranatha's claim that Jayabhadra preceded all 
the other commentators in his list gains support from the fact that Jayabhadra 
knew what is evidently the earlier of these two versions. It extends only to 50.19, 
ending with a passage on fire-sacrifices that may be performed if one wishes to 
subject another to one's will (vasyahomah). In the second, attested by all the 
other commentators except Bhavyakirti, 363 by the Tibetan translation, and by 

have seen, the Sarvabuddhasamayogadakinijalasamvara. 
363 In Bhavyakirti's Cakrasamvarapanjika the text of the Laghusamvara ends exactly 
where it does in Jayabhadra's. It is therefore likely to belong like Jayabhadra's to 
the earliest phase of the exegesis of this Tantra. Jayabhadra's appears to be the 
older of the two. In 41.8 Bhavyakirti attests with the later witnesses the interpo- 
lation (see here p. 199) *oddiyanajalandharapulliramalayadisu (bDe mchog nyung 
ngu, f. 239r2: au dya na \ dza la ndha ra dang pu li ra ma la ya sogs), since he 
comments here (f. 36v6): o dya na du ni 'od Idan ma'o | dza la ndha rar ni gtum 
mig ma'o | pu III ra ma la ya la sogs, whereas Jayabhadra says that Pulliramalaya 
has not been mentioned but must nonetheless be understood to be intended (p. 137: 
pulliramalayo na nirdistah sarvaplthanam pradhanatvad upadesad vavaseyah). It 
seems probable, then, that Bhavyakirti follows the reading of a subsequent redac- 
tion in which this 'omission' had been rectified. 

At the beginning of the translation the name of Bhavyakirti's commentary is said 
to be Suramanojna in Sanskrit and dpa' bo'i yid du 'ong in Tibetan, i.e. 'pleasing to 
heroes'. But the Sanskrit titles given in the Tenjur are so often inaccurate that we 
can conclude that they do not reach us from the Sanskrit works themselves but are 
reconstructions from the Tibetan added by the compilers of the Tenjur. The Sanskrit 
rendered by dPa' bo'i yid du 'ong can now only be guessed, but its first element was 
surely Vira- rather than Sura-. The Mahavyutpatti, composed to guide Tibetan 
translators and no doubt the dictionary used by the compilers of the Tenjur, gives 
dpa' bo to render both vira- and sura-, both meaning 'hero'; but though the two 


The Saiva Age 

the manuscript, the fiftieth chapter has eight verses after the last of the shorter 
text (50.20-27), followed by a fifty-first chapter of twenty-two verses. It is clear 
that the longer text is the later. For the alternative, that the shorter text arose 
after the longer by excision of the final thirty verses, is inconceivable, since these 
have the effect of greatly increasing the plausibility of the whole as a Buddhist 
work and were no doubt added because it was felt, quite rightly, that 1.1 to 
50.19 were inadequate in this regard. The only element of Mahayana Buddhist 
doctrine contained in the text up to 50.19 comprises a section of four verses 
(10.1-4) stating that success in the pursuit of Siddhis depends on the Sadhaka's 
identifying with the three Buddha bodies (Dharmakaya, Sambhogakaya, and 
Nirmanakaya), all other Buddhist elements being little more than a handful of 
occurrences of the terms Buddha, Tathagata, and Bodhisattva, and the names 
of Vajrayanist deities. 

Now Taranatha claims that his ten successive Tantric Acaryas of 
Vikramasila held their positions after the time of Buddhajnanapada and 
Diparikarabhadra, whom he places in the reign of Dharmapala (c. 775-812); 
and he reports that each did so for twelve years, implying thereby a form of 
limited tenure. Thereafter, he says, came the six "Door-keepers". Among them 
was Ratnakarasanti, who taught the Tibetan translator 'Brog mi Sakya ye 
shes (993-1077?) 364 and the Indian Diparikarasrijfiana 365 (982-1054), and was 
a slightly older contemporary of Jnanasrimitra, who was active c. 980-1030. 
From this it would be a simple matter to determine the approximate date of 
Jayabhadra, the first of the ten, by counting the years from either end, were it 
not that Taranatha makes the collective tenure of the ten Acaryas 120 years, 
whereas the interval between Diparikarabhadra and Ratnakarasanti is almost 
two centuries. We might be inclined to count back from Ratnakarasanti rather 
than forward from Diparikarabhadra, thinking that a historian's information 
is likely to be more reliable the closer he approaches his own time. In that 
case, if we trust Taranatha and set the end of the tenure of Kamalaraksita in 
1000, as the immediate predecessor of the Door-keepers, we will conclude that 
Jayabhadra's tenure ran from 880-892. 

words are synonymous in ordinary usage, in the tradition of the Yogimtantras it 
is the former alone that is used in the special sense evidently intended here, that 
is, as a technical term for the Tantric practitioner. As for the second element, the 
same dictionary gives manojna- for yid du 'ong. But the result is unattractive by 
the standards of Sanskrit authors, who generally sought, like authors everywhere, 
to give their works titles that appealed to the ear. Viramanoramd is synonymous 
and meets this requirement. 

364 Zhib mo rdo rje, p. 84. 

365 Blue Annals, p. 380. 


Genesis and Development of Tantrism 

However, this chronology can be reconciled with other reports only at a great 
stretch, at least for the later teachers in Taranatha's succession. Thus Dmar 
ston, pupil of Sa skya Pandita Kun dga' rgyal mtshan (1182-1251) tells us 366 
that Durjayacandra, who by this calculation would have held office from 940 
to 952, was the teacher of Prajnendraruci, also called Viravajra, and that the 
latter taught 'Brog mi Sakya ye shes. Now 'Brog mi is said to have let Tibet 
for Nepal and India when Rin chen bzang po was nearly fifty years old, 367 that 
is to say around 1007 if Rin chen bzang po was born in 958, as his biography 
claims and Gzhon nu dpal accepts, 368 and then, after spending one year in Nepal 
with Santibhadra 369 and eight at Vikramasila with Ratnakarasanti, 370 to have 
studied with Prajnendraruci for three or four, 371 that is to say, therefore, c. 1016- 
1020. If we accept that Durjayacandra is unlikely to have held such a senior 
post as that of the head Vajracarya of Vikramasila in his youth and assume for 
the sake of argument that he was fifty-five when he began his tenure, then if 
that tenure began in 940, he would have to have been continuing to teach long 
after his retirement at sixty-seven in 952, and Prajnendraruci, if we take 945 as 
the latest plausible year of his birth, would have been about seventy when he 
accepted 'Brog mi as his pupil. 

This scenario is not impossible; but neither is it comfortable. Nor is it helped 
by the fact that Prajnendraruci is reported to have collaborated with 'Brog mi 
on translations of texts pertaining to Hevajra and his consort Nairatmya. This 
evidence is given in the colophons at the end of these translations 372 and should 
be considered more reliable than that of hagiographical biographies. 

Even more difficult to reconcile is the report in the Chos 'byung of Pad ma 
dkar po (1527-1592) that Durjayacandra taught the Mantranaya at Vikramasila 
to the translator Rin chen bzang po. 373 For Rin chen bzang po is said to have left 
for India in 975, at the age of seventeen, and to have gone to Vikramasila only 
after a period of some seven years of education in Kashmir, therefore around 
982. At that time Durjayacandra would have been nearly a hundred if we hold 
to the assumption that he began his tenure in 940 when he was fifty-five years 
of age. 374 It is probable, then, that while we are indeed closer to the truth if 

366 Zhib mo rdo rje, pp. 86-88. 

367 Blue Annals, p. 205, 11. 26-31. 

368 Blue Annals, p. 68, 11. 3-6. 

369 Zhib mo rdo rje, p. 84, 11. 6-10; Blue Annals, p. 205, 11. 32-35. 

370 Zhib mo rdo rje, p. 86, 1. 10; Blue Annals, p. 206, 11. 18-19. 

371 Blue Annals, p. 206, 11. 32-33 (three years); Zhib mo rdo rje, p. 88, 11. 7-8 (four 

372 Toh. 1185, 1236, 1251, 1310. 

373 Tucci 1988, p. 35. 

374 Tucci 1988, pp. 3-4. 


The Saiva Age 

we calculate back from the Door-keepers than forward from Buddhajnana and 
Diparikarabhadra, Taranatha has placed the later teachers from Durjayacandra 
onwards too early. 

This suspicion gains further support from what we know of the life of 
Tathagataraksita. If Taranatha's report were accurate, provided that we 
calculate backwards from the six Door-keepers, then he would have held office 
at Vikramasila c. 964-976. But we learn from the colophon of the Tibetan 
translation of his commentary on the Yoginlsamcara that he translated the work 
himself with the help of the Tibetan Ba ri Lo tsa ba Rin chen grags. This places 
his activity well into the second half of the eleventh century. For Ba ri Rin chen 
grags is said by Gzhon nu dpal to have been born in 1040. 375 

If Durjayacandra, as now seems probable, was active towards the end of 
the tenth century, and if Taranatha is correct that there were no intervals 
between the tenures of his predecessors Jayabhadra, Sridhara, Bhavabhadra, 
Bhavyakirti, and Lilavajra, then we shall not be far from the truth if we assign 
them all these commentators on the Laghusamvara to the tenth century. 

Beyond the terminus provided by this tentative dating of the earliest 
commentators we have no clear knowledge of the date of these Tantras. It 
has been claimed by DAVIDSON that the Laghusamvara was already in ex- 
istence in the eighth century since Vilasavajra cites it several times in his 
commentary on the Manjusrinamasamgiti? 16 and this view has recently been 
repeated by GRAY. 377 The latter recognized that most of the former's claimed 
citations are actually not of the Laghusamvaratantra but of the Sarvabud- 
dhasamayogadakinijalasamvara, which Vilasavajra cites as the Samvaratantra, 
using the common abbreviation of this unwieldy title. But he argues that the 
date is established nonetheless by two places in the same commentary in 
which Vilasavajra cites a Cakrasamvaratantra or Cakrasamvaratantra. This 
GRAY takes to be the Laghusamvara under its commonly used alias. Both 
citations occur in a section of the commentary in which, explaining epithets 
found in the Manjusrinamasamgiti, Vilasavajra follows each with iti and the 
name of a Tantra in the locative, indicating that the epithet is also found 
in that source. The first citation, GRAY claims, is of Laghusamvara 2.16c (f. 
2v6: hasticarmavaruddham ca 'and [his back] covered with the hide of an 
elephant'), and the second of 48.12a (f. 35r6: kahkala mahakahkala). In fact 
the first passage does not cite Laghusamvara 2.16c, the text quoted being 
gajacarmapatardradhrk 'wearing as his upper garment the moist hide of an 

375 Blue Annals, p. 211. 

376 Davidson 1981, pp. 7-8. 

377 Gray 2007, pp. 12-14. 


Genesis and Development of Tantrism 

elephant', to which Laghusamvara 2.16c corresponds only in sense and then 
not exactly. 378 As for the second citation, 379 the word kahkalah does appear in 
the Laghusamvara, as the name of one of the twenty-four Vira consorts of the 
twenty-four Dakinis, 380 but as a single word its presence is not enough to es- 
tablish the identity of Vilasavajra's Cakrasamvara with the Laghusamvara. On 
the other hand, the fact that the first epithet attributed to the Cakrasamvara 
does not occur in the Laghusamvara is not sufficient to prove the opposing 
thesis, that Vilasavajra was referring to another work. For it is conceivable 
that he was citing the text not for the exact wording of Manjusrinamasamgiti 
69d (gajacarmapatardradhrk) but only for an expression close to it in meaning. 
But if this is true it establishes, of course, only that Vilasavajra may have been 
referring to Laghusamvara 2.16c, not that he was. To continue to hold to the 
position that Vilasavajra must have been referring to our Laghusamvara in 
spite of these considerations, one has to put one's trust in the fact that the 
Laghusamvara is also known as the Cakrasamvara and the fact that no other 
work of this name is cited (unless it be here). One must also remain free of 
the suspicion that there might have been another, earlier work with this title 
among the numerous Tantras known in the eighth century that have failed to 
survive either in Sanskrit or in Tibetan translation. 381 One must also overlook 
the evidence of the Laghusamvara itself. For that refers to a Cakrasamvara in a 
list of its own predecessors. 382 I conclude, therefore, that there is no more than 

378 Vilasavajra, Namamantrarthavalokinl A f. 57rl-2, on Manjusrinamasamgiti 69d 
(gajacarmapatardradhrk): gajacarmapatardradhrg iti sricakrasamvare | gajasya 
carma gajacarma patas casav ardras ca | gajacarmaiva patardrah gajacarmapata- 
rdrah | tarn dharayatiti gajacarmapatardradhrk. This error has been pointed out 
by Szanto (2008b, p. 217). 

379 Vilasavajra, Namamantrarthavalokinl A f. 55v6, on Manjusrinamasamgiti 67cd 
(damstrakaralah kahkalo halahalah satananah): kahkala iti sricakrasamvare. 

380 Laghusamvara f. 35r4-7 (48.9c-12): vajrasattva vairocana padmanartesvaras 
tatha | srivajraherukas caiva akasagarbha hayagrivam eva ca || 10 rat- 
navajra mahabala virupaksa bhairavas tatha | vajrabhadra subhadras caiva 
<va>jrahumkaram eva ca || 11 mahavira vajrajatilam tu ahkurika va- 
jradehaka \ vajraprabha amitabhah suravairino vikatadamstrinam eva ca \\ 

12 kahkala mahakahkala khandakapalinadi tu caturvimsativiranam sarvam 
vyaptam akhilamjagat. 

381 Such works cited in Vilasavajra's commentary are the Krodhendutilaka (A f. 
57r5), the Guhyakosa (A f. 57vl), the Vajraghanoccaya (B f. 39r6), the Satprajna- 
nayasamvara (B f. 40v3), the Sarvatantrasamuccaya (A f. 57r4), and the Va- 
jrakiriti (A f. 56v6). Similarly, in the Tattvasiddhi of Santaraksita we find the 
Sarvadevasamagama, the Laukikalokottaravajra, and the Vimuktisamudghatana, 
and in the Caryamelapakapradipa of Aryadeva the Vajramukhimahayoga and the 
Vi nayamoghas idd h i . 

382 Laghusamvara 27.23-24a as transmitted in Abhidhanottara, Patala 43, A f. 
140rl-2f, B f. 180v3-4: tattvasamgrahe yad uktam ca tathoktam cakrasamvare 


The Saiva Age 

a possibility that Vilasavajra knew the Laghusamvara and, therefore, that the 
existence of this Tantra in the eighth century remains unproved. 

What we can say with confidence is that the Laghusamvara came af- 
ter the Paramadya, the Vajrabhairava, the Sarvatathagatasamgraha, the 
Guhyasamaja, and the Sarvabuddhasamayogadakinijalasamvara, since it 
names these, 383 and tacitly incorporates verses from the last three in its earliest 
accessible redaction. 384 These borrowings do not rule out the possibility that the 

| guhyatantre samdkhydtam haritantre tathaiva ca || mahdbhairavatantre ca 
japavratddisiddhidam \ tad idam dhydnamdtrena mantri sddhayate ksandt. The 
reading cakrasamvare (cakrasamvare Cod.) is confirmed by Bhavabhatta in his 
commentary on this verse (Cakrasamvarapanjikd, p. 495). 

383 Laghusamvara f. 4v2-3 (3.22): abhisikto bhavet tatra sarvvatantraikam ut- 
taram | tattvasamgrahe samvare vdpi guhye va vajrabhairave; and f. 23v7 
(30.24): vidydrdjacakravarti ayam mantro na bhuyo na bhavisyati \ tattvasamgrahe 
paramddye samvare guhye va vajrabhairave. The Samvara here is the Sarva- 
buddhasamdyogaddkinijdlasamvara. The title is commonly so abbreviated; see 
also Indrabhuti's comment on the first passage ('Khor lo sdom pa'i rgyud kyi 
rgyal po bde mchog bsdus pa zhes bya ba'i mam par bshad, f. 38r7): bde 
mchog ni sgyu ma bde mchog go 'The Samvara is the Jdlasamvara' . In his 
Cakrasamvaravrtti Indrabhuti takes the Guhya here to be the Guhyasamaja or 
the GuhyendumanitilakalGuhyendutilaka (Toh. 477) (f. 38r7): gsang ba ni 'dus pa 
'am zla gsang thig le'i nor bu'i rgyal po'o. In his Cakrasamvaratikd Devagupta 
takes it to be "the Guhyasamaja etc." (f. 80r5): bsdus pa la sogs par. But in his 
Cakrasamvarapanjikd Bhavabhatta glosses guhyatantre in 27.23 as guhyakosddau 
'in the Guhyakosa etc.'. 

384 (1) Laghusamvara (LS) f. Iv5 (1.7c-8b): antargatena manasd kdmasiddhim tu 
bhdvayet | svaretobindubhir buddhdn bodhisattvdms ca pujayet < Sarvatathdgata- 
tattvasamgraha, section 2651: antargatena manasd kdmasuddhim tu bhdvayan | 
svaretobindubhir buddhdn pujayan siddhim dpnuydt, but influenced in the second 
line by Guhyasamaja 7.26: svavajram padmasamyuktam dvayendriyaprayogatah | 
svaretobindubhir buddhdn vajrasattvdms ca pujayet; (2) bDe mchog nyung ngu, f. 
234r5-6 (LS 31.1): de nas sha chen thams cad kyi | jigs byed rdo rje skyes yin bshad 
| 'di ni gdug pa thams cad kyi | jigs byed mi bzad par bshad do < Guhyasamaja 
5.78: mahdmdmsena sarvesdm ndsanam vajrajam smrtam | eso hi sarvakrurdndm 
ndsako ddrunah smrtah; (3) bDe mchog nyung ngu, f. 234v4 (LS 31.12): sa ni 
spyan zhes bya bar bshad | chu khams ma ma ki ru brjod \ me ni gos dkar mor 
bshad de | rlung ni sgrol mar rab tu brjod < Guhyasamaja 17.51: prthivT locand 
khydtd abdhdtur mdmakl smrtd | pdndardkhyd bhavet tejo vdyus tdrd praklrtitd; 
(4) LS f. lv (1.1-3) < SarvabuddhasamdyogaddkinTjdlasamvara (SBSDJS) 1.1-2 
etc. (see here p. 154); (5) LS f. lv5-6 (1.8c-9b): darsanasparsandbhydm ca sravane 
smaranena ca || mucyate sarvapdpais tu evam eva na samsayah < SBSDJS as 
quoted in Jndnasiddhi 15.50: darsanasparsandbhydm ca sravanasmaranena ca \ 
sarvapdpair vimucyante *yujyante (em. : pujyante Ed.) sarvasiddhibhih (= Sangs 
rgyas thams cad mnyam par sbyor ba f. 152v3 [2.16]); (6) LS ff. Iv7-2rl ( 
madhu raktam sakarpuram raktacandanayojitam | ganamadhye pratisthan tu [+ 
sarvocchistarasdyanam in the earlier redaction incorporated in the Abhidhdnottara 
A f. 146rl— 3 (46.3— 5b)] sarvavajrdhkacihnadhrk \ andmdhgusthavaktrdbhydm 
lehayed yogavit sadd || somapdnavad dsvddya siddhim dpnoti sdsvatlm < Sangs 
rgyas thams cad mnyam par sbyor ba f. 158v4-5 (SBSDJS 6.15—17): dmar 


Genesis and Development of Tantrism 

Laghusamvara was composed in that century, since none of the works is later 
than that time. But three considerations suggest a later date. (1) No text of the 
Cakrasamvara corpus, or any other Yoginitantra, was translated into Tibetan 
during the earlier diffusion of Buddhism (snga dar) that occurred from the 
eighth century to the middle of the ninth, during Tibet's imperial period: this 
new literature reached the Tibetans only during the later transmission (phyi 
dar), that began c. 1000. (2) Among the many surviving stone, metalwork, and 
painted Indian images of Samvara none is demonstrably earlier than the tenth 
century 385 Finally (3), there is, as we have seen, no evidence of commentatorial 
work on the Laghusamvara before c. 900. Of course, none of these facts proves 
conclusively that the Laghusamvara was not in existence at an earlier date. But 
they do incline one to consider a later date more probable. This is particularly so 
in the case of the absence of commentaries. The Laghusamvara is so problematic 
text from the Buddhist point of view that it is hard to imagine that it could have 
survived for long without the support of learned exegesis. 

Whatever its date, the Laghusamvara is likely to be a product of the first 
phase of the development of the Yoginitantras, if not the earliest of them all. 
This surmise rests on the assumption that Yoginitantras that are less sophis- 
ticated in the sense that they show a less developed Mahayana Buddhist theo- 

chen dang ni ga bur bcas \ tsa ndan dmar por sbyar ba dag \ tshogs kyi 
dbus su bzhag pa ni | ra sa ya na kun slong ba | rang gi lha yo sbyor Idan 
pas | srin lag dang ni mthe bo'i rtses | zhi ba'i btung pa bzhin myangs na 
| Hag pa yi ni dngos grub thob (*mahdraktam sakarpuram raktacandanayoji- 
tam | svddhidaivatayogena sarvocchistarasdyanam | andmdhgusthavaktrdbhydm 
<++++++++> | somapdnavad dsvddya siddhim apnoti sdsvatlm); and (7) LS f. 12r6- 
7 (13.2): yadyad indriyama.rgatvamya.ydt tat tat svabhavatah | paramdhitayogena 
sarvam buddhamayam vahet <SBSDJS as quoted in Carydmeldpakapradipa, p. 90: 
yad yad indriyamdrgatvam ydydt tat tat svabhavatah | asamdhitayogena sarvabud- 
dhamayam vahet. 
385 A Kashmirian Samvara of leaded brass inlaid with copper and silver in the Los 
Angeles County Museum of Art from the Nasli and Alice Heeramaneck Collection 
appears to have been assigned by Pal to c. 725 in his catalogue of the exhibition 
'The Arts of Kashmir' (2007, p. 91, fig. 92). However, he has kindly informed me 
(personal communication, 1 March, 2008) that this surprisingly early date is not 
his own but that of the museum (for which see 
recorded on the loan agreement form. The lending museum insisted on this date 
and it was substituted for his own without consulting him. He had assigned it to "ca. 
9th century". In an earlier publication (1975, p. 173, pis. 64a,b) he had proposed the 
tenth. Reedy (1997, p. 162, fig. K62) gives '9th-10th century'. Linrothe (1999, p. 
289, fig. 211) has found these dates too early and suggests the late tenth or early 
eleventh century. In the absence of a detailed art-historical demonstration of the 
date, which I suspect could in any case be no more than tentative given the small 
population of comparable pieces, I am inclined in the light of the other historical 
evidence to agree with Linrothe. 


The Saiva Age 

retical framework are likely to be earlier than those in which the level of the- 
oretical assimilation is more advanced. By this criterion the Hevajra must be 
placed after the Laghusamvara. This also assumes that the development of the 
Mantranaya was not unilinear throughout, since if it were we would have to 
place the Laghusamvara before the Mahavairocanabhisambodhi, Sarvatatha- aha, Guhyasamaja, and Sarvabuddhasamayoga. It assumes, 
then, that the Yoginltantras represent a new phase with its own humble begin- 
ning, and that it was only later in this phase that the tradition got up to speed, 
as it were, by fully integrating the new world of practice whose entry marks its 
commencement by providing it with a thoroughly Buddhist encoding. While it 
is possible that this assimilation of the text began long after its first redaction 
it seems more probable in the absence of firm evidence to the contrary that if 
so problematic a creation were to have remained for long without the benefit of 
learned exegesis it would be likely to have disappeared without trace. 

As for the provenance of the Laghusamvara, it was certainly eastern India, 
the region in which most of the Indian learned exegesis of this Tantric corpus 
was produced. The Tantra does not state this explicitly. Claiming the status 
of revelation it would have been averse to doing so. Nonetheless, it reveals 
its provenance in spite of itself by giving BA in its encoding of some of the 
syllables of Mantras where correct Sanskrit requires VA. This is evidently an 
effect of the fact that va is pronounced ba in the Indo-Aryan vernaculars of this 
region. 386 Thus 5.4 yields BHAGABATE rather than BHAGAVATE: pancamasya 
yac caturtham prathamasya trtlyam \ trayovimsas tathaiva ca caturthasya yah 
prathamam (f. 5r3-4) 'the fourth of the fifth [class of consonants] (BHA), the 
third of the first (GA), the twenty-third (BA), and the first of the fourth (T-)' ; and 
30.20-21 yields BHAGABAM rather than BHAGAVAM (for BHAGAVAN): kosthakad 
dasamam caiva vilomena tu sadhakah \ kosthaka ekonavimsatimam tatha 
trayovimsatikosthakad \ dvitlyakosthasamyuktam binduna urdhvabhusitam (f. 
23v4-5) 'The Sadhaka should select the tenth counting backwards from the 
compartment [of HA] (BHA), the nineteenth from that [of A] (GA), and the [letter] 
from the twenty-third box (B-) together with [the letter in] the second box 
adorned above with a dot (AM)'. 

Variant readings giving the correct spellings in these cases are found. In 5.4 
Jayabhadra and Bhavabhatta read ekonnatrimsa- (sic) and ekonatrimsati 'the 
twenty-ninth' (VA) rather than the trayovimsa- 'the twenty-third' (BA) seen in 
the Baroda manuscript; and this reading is also found in the Tibetan translation 
(de bzhin nyi shu tsa dgu la [- ekonatrimsam tathaiva ca]) and the redaction 

386 This is so in Bihari, Maithih, Bengali, Kumauni, Nepali, Assamese, and Oriya. 


Genesis and Development of Tantrism 

of this passage in Patala 54 of the Abhidhdnottara (A f. 166r3: ekonnatrinsam 
tathaiva ca). Likewise in 30.21 we find Jayabhadra giving ekonatrimsati- (VA) in 
place of the reading trayovimsati (BA) attested by the manuscript, but here the 
'incorrect' reading is also supported by the Tibetan translation and the commen- 
tary of Bhavabhatta. There can be little doubt that the non-standard readings 
giving BA rather than VA are original. For it is not surprising there should have 
been attempts to correct an original BA to VA, whereas it would be most unlikely 
that any redactor would have made the effort to rewrite a reading that gave VA 
in order to yield BA. 387 

Also indicative of the east-Indian provenance and development of this cor- 
pus are the form chamdoha- in place of samdoha- , 388 and the pervasive promis- 
cuity of the forms -samvara- and -samvara- in the names of its deity, in the 
title of the primary Tantra, and in the compound in which this form is pre- 
ceded by dakinljala- or yoginljala- . I use the forms Samvara and Cakrasamvara. 
Laghusamvara and Cakrasamvara here in keeping with the usual Tibetan trans- 
lations, namely bDe mchog and 'Khor lo sdom pa; and this accords with se- 
mantic analyses of these names and titles in the Sanskrit commentators. Thus 
Bhavabhatta explains the second element of the second in the sense 'he who re- 
strains' from the the verb samvr-, and construes the whole to mean 'he who by 
means of the wheel (cakra-) [of the Dharma] restrains [the minds of living be- 
ings from the wrong path] (-samvarahf (*cakrena samvrnotiti cakrasamvarah), 
telling us further that the name is extended to the Tantra because this de- 
ity is its subject. 389 As for the form Samvara, that too is widely supported. 

387 It is not probable that the Laghusamvara was alone among the Yoginitantras in 
being of east-Indian origin. We see the same tell-tale B- for v- in 1.4.27-28 of the 
Catuspltha, the Mantra syllables VADAVE being encoded there as BADABE. More- 
over, it is probable that the Apabhramsa seen in some verses of the Hevajra is of the 
eastern variety. This is suggesred by the nom. sg. endings -oho and -aha in kibidaho 
in 2.4.6 and hutasanaha in 2.4.67; see Tagare 1987, p. 110-111. An investigation 
of the language of the Apabhramsa verses that are found in such Yoginitantras as 
the Hevajra, Khasama, Catuspltha, and Dakarnava, in comparison with that of the 
Doha collections of Kanha and Saraha, may be expected to shed more light on this 
question of provenance. 

388 See here p. 180. 

389 Bhavabhatta, Cakrasamvarapanjikd, explaining the title with the prefixed honoric 
Sri- when it occurs in the final colophon in the words sricakrasamvaranamni raaha- 
yoginitantraraje 'in the great king among the Yoginitantras called srlcakrasamvara' 
in the final colophon: srih punyajnanasambharah \ cakram dharmacakram | srimac 
cakram srlcakram | tena kapathat sattvanam manah samvrnotiti srlcakrasamvarah 
sriherukah | tadabhidhayitvat tantram api tathocyate 'The word Sri- 'glory' denotes 
the accumulating of [bothl merit and gnosis. The word -cakra- 'wheel' refers to the 
wheel of the Dharma. It is prefixed by Sri to express the fact that it [, that is to 
say, the teaching of the Buddha,] entails this [provisioning with both merit and 


The Saiva Age 

Ratnakarasanti explains it as meaning 'the Highest (varam) Bliss' (sam) when 
analysing its occurrence in the neuter in the compound dakinijalasamvaram; 390 
and Bhavabhatta when analysing its occurrence in the masculine gender at the 
end of the same (dakinijalasamvarah) takes it to mean '[Heruka,] who protects 
Bliss (sam vrnotlti samvarah) [by keeping it free of all defects]'. 391 This line of 
analysis, which applies a meaning of sam that is well-attested in non-sectarian 
lexicography 392 is not the invention of these commentators. They draw on the 
authority of the Sarvabuddhasamayoga, which refers to its deity Vajrasattva 
as Samvara and explains that name as meaning '[he who has/is the] Highest 
Bliss'. 393 That the -samvara form is not only old but also original is established 

gnosis]. Heruka is called Sricakrasamvara [here] because he restrains [samvrnotiti 
samvarah] by means of this [wheel, in the sense that he restrains] the minds of 
living beings from the false path. [This] Tantra has the same name because it is 
that which refers to him'. 

390 Ratnakarasanti, Mahamayatlka on 23d: samvaram sukhavaram mahasukham 
'[sam means 'bliss' and -varam 'best'. So] samvaram means 'the best bliss' 
(sukhavaram) [, i.e.] 'the Great Bliss' (mahasukham)'. The same analysis is tacitly 
given in such parallel expressions as dakinijalasatsukham in Samvarodaya 3.6d 
and 26.10cd; and Vajraddka l.lcd: sarvadakinlmayah sattvo vajraddkah param 
sukham; 1.12cd, 1.50,1.71cd: sarvaddkinisamdyogavajraddkah param sukham. 

391 Bhavabhatta, Cakrasamvarapahjika on 1.2: dakinl sunyata. jalam upayah \jdlena 
hi matsyadibandhanasiddhih \ updyena hi klesamlnddir niyamydkimcitkarah kriy- 
ate | tabhyam sam sukham avadyebhyo bahiskrtya vrnotlti dakinijalasamvarah 
'[The meaning of the name] Dakimjalasamvara [applied to Heruka here] is 'he who 
protects (-varah [vrnotlti varah]) bliss (sam) by means of the Dakim and the Net 
(jalam)'. The term Dakini [here] means ['Emptiness',] 'the fact that [all things] are 
void of [intrinsic reality]' (sunyata); and the term 'Net' refers to the method (upayah) 
[, namely the compassion (karuna) that must accompany awareness of that Empti- 
ness]. It is called a net [metaphorically]. For by using a net one succeeds in catching 
fish and other creatures. [Likewise] by employing the method [that is compassion] 
one restrains and so renders incapable of activity the 'fish and other creatures' that 
are the afflictions (klesdh)[, namely attachment, hatred and the rest]. He protects 
bliss by means of these two[, emptiness and compassion,] in the sense that through 
these he protects it from [those] defects'. 

392 See, e.g., Hemacandra, Anekdrthasamgraha, Parisistakdnda 21a: sam kalydne 
sukhe 'tha; Vardhamana, Ganaratnamahodadhivrtti, p. 39, on 1.15: sam 
duhkhopasame; Yaska, Nighantubhdsya, p. 521 (on Rgveda 5.4.5: sam no bhavantu 
vdjinah): sukhdh no bhavantu vdjinah. 

393 Sangs rgyas thams cad mnyam par sbyor ba, f. 154r6-7 (1.10): sham zhes bya ba 
bde bar bshad | sangs rgyas kun gyi bde chen yin | sgyu ma thams cad rab sbyor 
ba | mchog tu bde bas bde ba'i mchog (sukham sam iti vikhydtam sarvabauddham 
mahasukham \ sarvajdlasamdyogah sukhavarena samvarah) 'The word sam means 
'bliss', the Great Bliss of all the Buddhas. He is Samvara because of [the fact that 
he possesses] the highest degree of [this] bliss'. The Sanskrit of the first half of 
this verse is supported by its citation by Vilasavajra while explaining the epithet 
mahdsukhah in his Ndmamantrdrthdvalokini, f. 57vl— 2: mahdsukha iti srisamvare 

| tatra mahdsukha iti yat tdthdgatam andsravam sukham tan mahdsukha ity u- 
cyate \ tatraivoktam sukham sam iti vikhydtam sarvabauddham mahasukham iti. 


Genesis and Development of Tantrism 

by evidence outside the Buddhist corpus. For Yoginljalasamvara is found as the 
name of a Bhairava in one of the secondary Kalpas taught in the Sakta Saiva 
Picumata, 394 which, as we shall see, was a major unacknowledged source for 
the redactor of the Laghusamvara. That the form intended there is -samvara- 
rather than -samvara- is certain, because the text provides a semantic analysis 
that takes the first syllable to mean bliss (sukham). 395 

However, these are not the only views. Jayabhadra, commenting on 1.2 of 
the Laghusamvara, takes the same expression to be Dakinijalasamvara, un- 
derstanding it to refer to the Laghusamvara itself and explaining it as 'The 
Concealment of the Array of Dakinis', deriving the last element of the com- 
pound from samvr- 'to envelop'; 396 and while the Tibetans usually render the 

The first Pada is also supported by Bhavabhatta, who quotes it without attribu- 
tion, when explaining dakinijalasamvaram in Laghusamvara 1.2: sam sukham iti 
cakhydtam iti vacanat. 

394 Yoginljalasamvara in this text is a form of Bhairava and the term refers by 
extension to his Mantra and the associated system of practice (vratam). See 
Picumata f. 251r5-vl (56.4c-6b): srnu devi pravaksyami sarvayogiprasadhanam 
| yagamantrasamopetam yoginljalasamvaram | yena vijnatamatrena trailokye khe- 
caripadam \ asadya kridate mantri kulasiddhisamanvitah 'Listen, O Devi. I shall 
teach you about Yoginljalasamvara together with the deities with whom he is to be 
worshipped iyaga-) and his Mantra, as the means of propitiating the Yoginis. As 
soon as the Mantra adept has mastered this he will reach the domain of the Khe- 
caris and move freely through the triple universe, possessing [all] the supernatural 
powers of the [Yogini] clans'. 

395 Picumata f. 251v2— 3 (56.12— 13b): samuham jalam ity uktam yogininam maho- 
dayam \ sam sukham vara datrtva<t> *samuhatvavivaksaya (samuhatva em. : 
samuhatvam Cod.) | * yogesiyogabhdvastham (yogesiyoga conj. : yogayogisa Cod.) 
yoginljalasamvaram | mantram tu kathitam devi bhairavasyamitatmakam 'The ex- 
pression Yogimjala [in Yoginljalasamvara] means the exalted totality of the Yoginis, 
jalam 'net' denoting 'multitude' [here]. The sam of -samvara means 'bliss' (sukham). 
The YoginIjalasamvara[mantra] is so named because it is the bestower {-vara) of 
that bliss, [-vara- being formed as an agent noun from the verb vr- 'to give']. It is 
the granter of this bliss to the Yogimjala in as much as it is located in the inner state 
of *the Yoga of the Yogesvaris, the plurality of these being intended in the sense of 
their totality (conj.). The Mantra of Bhairava [that bears this name] is infinite [in 
its power]'. 

396 Jayabhadra, Cakrasamvarapahjika on 1.1— 2b (athato rahasyam vaksye samasan na 
tu vistarat | sriherukasamyogam sarvakamarthasadhakam | uttarad api cottaram 
dakinijalasamvaram 'Next I shall teach the secret, in brief rather than at length, 
the congress of Sriheruka, the accomplisher of all desires, the Dakinijalasamvara, 
higher even than the higher'): uttarad api cottaram iti desyadesakayor abhedat | 
yany uttaratantrani samajadini tesam apy uttaratvad uktam \ dakinijalasamvaram 
iti | dakinyah sarvas tricakravyavasthitdh \ tasam jalah samuhas tasya samvarah \ 
samvaranam gopanam ity arthah 'It is referred to as higher even than the higher 
because it is higher even than the Tantras [of the Yogottara class] headed by the 
[Guhya]samdja, which are 'higher' because the difference between teacher and the 
taught is absent [in them]. As for [the title] Dakinijalasamvara, it means the con- 
cealing of the net, that is to say, of the totality of all the Dakinis that are established 


The Saiva Age 

Cakra- name 'Khor lo sdom pa and so support the form Cakrasamvara, we also 
find 'Khor lo bde mchog in their translations, which supports the alternative 
C akras amvara. 397 

The reason for this inconstancy is evidently that sa and sa are both pro- 
nounced as sa in Bengali, as they were in the Magadhi Prakrit of the drama- 
tists. 398 Consequently, instead of attempting to decide which form is correct we 
should recognize that for the east-Indian followers of this tradition there was in 
effect only one word here (samvara / samvara), which could be understood either 
as 'the highest {-vara- [Tib. mchog]) bliss (sam [Tib. bole])' or as 'fusion' and the 
like by derivation from the verbal root vr preceded by the preverb sam. That this 
was the case is demonstrated by a passage in the Samvarodaya in which the two 
semantic analyses, explaining samvara- and samvara- respectively, are given for 
one and the same word. 399 

Samvara/Vajrarudra and VajravarahI: The Transformation of 
BHAIRAVA AND HIS CONSORT. What marks the new start seen in the Yogini- 
tantras is a far more comprehensive adoption of the practices of the Saiva 
Vidyapitha texts, to the extent that there is little in the observances of these 
texts that does not draw on that source. Heruka is now paired with a lustful 
consort (VajravarahI in the Cakrasamvara texts and Nairatmya in those of 
Hevajra), and in the case of the Cakrasamvara tradition, so are the principal 
Yoginis of his retinue, a feature that matches the practice of the Vidyapitha's 
Picumata (Brahmayamala). Moreover, in the case of the tradition elaborated on 
the basis of the Laghusamvara the icon of Heruka has several blatantly obvious 
features of the iconography of Siva (/Bhairava) in addition to those manifest in 

in the three circuits [of the Mandala of Cakrasamvara], samvarah being derived 
from the verb samvr- 'to conceal' in the sense of the action of concealing'. 

397 In the DT 'khor lo sdom pa (cakrasamvara-) occurs about 250 times and 'khor lo bde 
mchog (cakrasamvara-) about 100; see, e.g., DT, Rgyud 'grel, vol. cha, f. 242v3 ('khor 
lo bde mchog gi gzugs can); vol.ja, f. 58v7 ('khor lo bde mchog gi rgyud), and f. 102r7 
('khor lo bde mchog gi sngags). 

398 See, for Magadhi, Vararuci, Prakrtaprakasa 11.2: sasoh sah 's is used in place of 
both s and s'. Generally in Middle and New Indo-Aryan the three Sanskrit sibilants 
have been reduced to s. It has been reported that in the Tantric Buddhist Doha 
texts, composed in what has been called Eastern Apabhramsa, s has been preserved 
in derivatives of words that have it in Sanskrit (Tagare 1987, p. 77). It is true that 
a few such forms are found in the manuscripts (Shahidullah 1928, p. 37), but 
there are many cases in which s does not appear, such as sunna for Skr. sunya. It is 
likely that the occasional distinction between s and s was learned window-dressing 
and that both consonants were pronounced s. 

399 Samvarodaya 3.17c-19b: samvaram sarvabuddhanam evamkare pratisthitam || 
kayavakcetasam karma sarvakaraikasamvaram \ samvaram sukhavaram bodhir 
avacyam anidarsanam || rahasyam sarvabuddhanam milanam samvaram varam. 


Genesis and Development of Tantrism 

the Heruka of the Sarvabuddhasamdyoga. He is black-bodied, and has twelve 
arms and four faces, with three eyes in each. He stands in the warrior pose 
with a Vajra and a Vajra-topped bell in his two principal hands, holding the 
bleeding hide of a flayed elephant over his back with his two uppermost hands, 
and in the remaining eight a rattle-drum (damaruh), a battle-axe, a chopping 
knife, and a trident, a skull-topped staff (khatvdngah), a skull-bowl (kapdlam) 
filled with blood, a lasso (pdsah), and the severed head of the god Brahma, 
wearing a long garland of fifty bleeding human heads around his neck, adorned 
with five ornaments of human bone and the ash of cremation-pyres smeared 
over his limbs — these, the bone ornaments and ash, are the Six Mudras of the 
Kapalikas — , with a tiger skin around his waist, a brahmanical cord in the 
form of a snake (ndgayajnopavltah), and a chaplet of skulls (kapdlamdld) above 
his forehead, his hair arranged in a high crown-like mass of ascetic's braids 
(jatdmukutah) adorned at the front with two crossed Vajras (visvavajram) and 
the new moon. His consort Vajravarahi stands before him in sexual union, with 
Heruka holding her to his chest with the hands that hold the Vajra and the 
Vajra-bell crossed at the wrists behind her back. She is red, one-faced, and 
two-armed, naked but for a filigree of fragments of human bone adorning her 
hips (asthimekhald), her right arm raised aloft holding a chopping-knife, with 
her index finger extended in a gesture of threatening the wicked, and her left 
arm, wrapped around Heruka's neck, holding to their mouths a skull bowl full 
of human blood and entrails, wearing a garland of fifty desiccated heads and the 
five Kapalika bone-ornaments, laughing, and intoxicated by lust. They are sur- 
rounded by a retinue of thirty-six goddesses termed Yoginis, Dakinis, Viresvaris, 
or Virinis visualized in the same Kapalika style, in concentric circuits of four, 
twenty-four, and eight, the twenty-four embracing Vira consorts and worshipped 
as residing in twenty-four sacred sites covering the whole subcontinent, from 
Uddiyana in the north to Ramesvara at India's southern tip, from Sindhu in 
the west to Devikotta in the east. The whole is surrounded by a ring of eight 
cremation grounds. 400 

The features of Siva's iconography evident here are the trident, the third 
eye, the new moon on the piled up braids, the tiger-skin lower garment, the 
multiple faces and arms, the skull-bowl, the skull-staff, the bleeding elephant 
hide, the severed head of Brahma, the snake as brahmanical thread, the sharp 
fangs, the chaplet of skulls, his dwelling in the cremation grounds, and the ashes 

400 This description of Heruka and Vajravarahi follows that given by Jayabhadra in 
his Cakrasamvarapahjika, p. 109, on Laghusamvara 1.10. for the iconography of 
the Yoginis and Viras see Bhavabhatta's Cakrasamvaravivrti on Laghusamvara, 
Patala 4 (vol. 1, pp. 44-47). See also Nispannayogdvali, pp. 26-29. 


The Saiva Age 

on his limbs. All these had entered Siva's iconography long before the forma- 
tion of the Tantras of the Cakrasamvara cycle. Siva's trident appears on seals 
and intaglios during the Kusana and Kusano-Sassanian periods in Gandhara 
and Afghanistan. 401 The third eye appears in sculptures of Siva from Mathura 
around the beginning of the third century; and the ascetic's piled braids and 
the new moon upon them appear there and elsewhere from the beginning of the 
fifth; 402 and all these characteristics, the trident in his hand, the third eye, the 
ascetic's braids, and the new moon, are mentioned in the Mahabharata, 403 as 
are his tiger-skin, his multiple faces and arms, his skull-bowl, his skull-staff, his 
brahmanical thread in the form of a snake, his sharp fangs, his garland of skulls, 
and his living in the cremation grounds smeared with ashes from its funeral 
pyres. 404 His wearing a bleeding elephant hide is also a commonplace by that 
time, being mentioned along with his crematorial characteristics in the works of 
the poet Kalidasa. 405 As for the severed head of Brahma, this too derives from 
a well-known Saiva myth which though not found in the Mahabharata in the 
text common to all the regional versions, 406 does appear in the Skandapurana- 

401 For a recent analysis of Siva images in the subcontinent, including those on coins, 
from the first century B.C. to the end of the Kusana period, see Ghose 2002, pp. 

402 Kreisel 1986 (Mathura, c. 400), p. 82; Barker 1997, pp. 149-151 (Mansar, c. 

403 Mahabharata 3.8.111a (trisulapaneh); 13.14.119 (balendumukutam . . . tribhir ne- 
traih krtoddyotam), 12.122.24b (sulajatadharah), 7.172.59c (jatamandalacandra- 

404 See, e.g., Mahabharata 13.127.18a (vyaghracarmdmbaradharah); 14.8.30d (maha- 
devam caturmukham), 13.14.116c (astadasabhujam sthanum), 14.8.28a (viriipa- 
ksam dasabhujam), 13.17.40a (dasabahus tv animiso); 12.36.2c (kapalapanih 
khatvahgi), 10.7.4d (khatvahgadharinam); 13. 15. lied (tiksnadamstram ...vyala- 
yajhopavitam), 14.8.21a (tlksnadamstraya karalaya); 10.6.33c (kapalamalinam); 
10.7.4a (smasanavasinam); 13.14.153c (suklabhasmavaliptaya) . 

405 Meghadiita 36c: hara pasupater ardrandgajineccham 'Remove Siva's desire for his 
[blood-]wet elephant hide'; Kumarasambhava 5.67d: gajdjinam sonitabinduvarsi 
ca '[his] elephant hide that showers drops of blood'; 5.77b: trilokanathah 
pitrsadmagocarah 'The Lord of the Three Worlds frequents cremation grounds'; 
5.69c, 5.79b: citabhasmarajah 'the ash-dust of funeral pyres'; and 5.71b: kapdlinah 
'decked with skulls'. Rudra/Siva frequently has the epithet krttivasas- 'wearer of 
the hide' in the Mahabharata. The Matsyapurana (Patala 153) relates that this is 
the hide of the elephant demon Gajasura killed by Siva in a great battle between 
the gods and the Asuras. How the elephant hide was understood when incorporated 
into the iconography of Heruka is not stated in most instances of its mention. But 
in two Kalpas in the Abhidhanottara, those of Samayasamvara and the Heruka of 
the ekaviravidhanam, it is said to be that of the elephantine Saiva-brahmanical de- 
ity Ganapati (B f. 34vl: aparabhujadvayena ganapaticarmdmbara^dharam (corr. : 
dhara Cod.) and (B f. 40v2-3: aparabhujadvayena ganapaticarmambaradhar ah). 

406 There is a reference to it in a supplementary passage of 26 verses inserted within a 


Genesis and Development of Tantrism 

Ambikakhanda, 407 probably composed in the sixth or perhaps the first half of 
the seventh century. 408 Other features in addition to these, namely the garland 
of severed or desiccated heads, the chopping knife, the rattle-drum, the Kapalika 
bone-ornaments, the consort, the skull-bowl full of blood and entrails, the retinue 
of Yoginis, their pairing with Vira consorts, the sacred sites, the theriocephalic 
gate-guardians, and the encircling cremation grounds, are commonplaces of the 
iconography of the Vidyapitha texts. Only the Vajras place a Buddhist seal on 
the icon. 

The image, then, has every appearance of representing a Buddhist trans- 
formation of Siva himself in his Bhairava aspect. Indeed in his commentary on 
the Laghusamvara Jayabhadra refers to this Heruka as Vajrarudra, that is to 
say, as Siva/Bhairava converted and liberated by assimilation into the essence of 
Buddha-hood, 409 thereby definitively surrendering and transcending his Saiva 
identity. In clear expression of this transcendence Heruka/Vajrarudra and Va- 
jravarahi are depicted and visualized standing on the sprawling, terrified bodies 
of a black Bhairava and a red, emaciated Kalaratri, their own pre-Buddhist iden- 
tities as the principal deities of the Vidyapitha. 410 

hymn to Siva (13.14.150-166) after 13.14.153 in the Maithili and Bengali versions, 
the Devanagari version of the commentator Nilakantha, in several manuscripts 
of the composite version, and the Kumbhakonam edition (Anusdsanaparvan, Ap- 
pendix I, no. 6, 1. 45): brahmasiropahartdya '[obeisance] to the remover of Brahma's 

407 5.1-63 (ed. Adriaensen, Bakker, and Isaacson, pp. 132-141). 

408 See here p. 51. 

409 Jayabhadra, Cakrasamvarapanjikd on Patala 12: krtapurvasevo mantrirdt iti 
vajrarudrayogavdn 'When the king among Mantra adepts has completed the 
preparatory service (purvasevd), that is to say, when he has achieved a state of 
complete identification with Vajrarudra . . . '; and on Patala 27: jndnahetujam iti 

| jndnasya prakarsaparyantam | tasya hetuh kdranam bhagavdn vajrarudrah \ 
tasmdj jdto bhavatity arthah 'jnanahetujam means born from the cause of knowl- 
edge, where knowledge is wisdom's ultimate degree and its cause is Lord Vajra- 
rudra'. Vajrarudra appears already in the Sarvabuddhasamdyoga in a passage 
that associates the nine dramatic sentiments (rasdh) with Vajrasattva, Tathagata, 
Vajradhara, Lokesvara, Vajrasurya, Vajrarudra, Sakyamuni, Arali (or perhaps 
Aralli), and Sasvata (Vairocana) respectively. Vajrarudra's is the sentiment of ter- 
ror (bhaydnakarasah) and it is probable therefore that we should understand Vajra- 
rudra to be Heruka. Sangs rgyas thams cad mnyam par sbyor ba f. 128r3: rdo rje 
sems dpa' steg pa la | dpa' la dpa' bo de bzhin gshegs | rdo rje 'dzin pa snying rje la 
| rgod pa jig rten dbang phyug mchog | rdo rje nyi ma khro ba la | rdo rje drag po 
'jigs pa la | shd kya ihubpa mi sdug la | ngo mtshar la ni a rali\ rab tu zhi la sangs 
rgyas rtag C'srhgdre vajrasattva hi vire caiva tathdgatah | vajradhrk karundydm 
tu hdsye caiva lokesvarah | vajrasuryas tathd raudre vajrarudro bhayanake \ 
sdkyamunis tu bibhatse drallir adbhute tathd \prasdnte sdsvatas caiva). 

410 Kalaratri here is the fearsome emaciated goddess variously called Carca, Carcika, 
Camunda, and Karnamoti; see here p. 231. 


The Saiva Age 

The Rise of the Goddess to Independence. Here Heruka's consort is 
visibly his dependent: while he has four faces and twelve arms she has only one 
and two. But in the subsequent development of this tradition we find a strongly 
Sakta tendency to elevate her to equality with Heruka and eventually to superi- 
ority, just as occurred in the development of the Vidyapitha. 411 Thus in certain 
other Kalpas in which Heruka is united with Vajravarahi at the centre of the 
Mandala her status is raised by endowing her with four faces and four or more 
arms. This is the case in the Kalpa of the sixth Patala of the Abhidhdnottara, 
which teaches what it calls the ekaviravidhdnam, the procedure in which the two 
deities alone are worshipped as 'solitary heroes' (ekavira-), that is to say, without 
the the retinue of the thirty-six Yoginis and twenty-four Viras. Here Heruka has 
twelve arms and Vajravarahi four, holding a blood-filled skull-bowl, a chopping- 
knife raised aloft with the gesture of threat, a rattle-drum, and a skull-staff. But 
both have four faces. 412 In the seventh Patala a two-faced, six-armed Vajrasattva 
transforms into a six-faced, twelve-armed Heruka Manjuvajramahasukha ac- 
companied by a Vajravarahi who has the same number of faces and arms and 
holds the same attributes in her hands. Brahma's severed head is absent here, 
but Brahma himself is not: his flayed skin takes the place of the elephant hide; 
and in place of a tiger skin we see that of Bhairava. 413 We see the same equality 
in the tenth Patala, where both Heruka and Vajravarahi are five-faced and ten- 

411 See Sanderson 1988, pp. 668-678. 

412 Abhidhdnottara B f. 40r3: athdnya<m> sampravaksydmi ekaviravidhdnakam 
| . . . (f. 40r6) sriherukam dtmdnam bhdvayet \ caturmukham dvadasabhujam 
. . . (f. 41rl-3) tasydgrato dlikdlisthitd bhagavati vajravarahi raktavarnd catur- 
vaktrd caturbhujd trinetrd muktakesi | nagnd khandamanditamekhald | vdme 
bhujdlihganakapdlam ca dustamdrddyasrgbodhicittaparipurnam daksine tar- 
janivajrakartikd | aparabhujadvaye damarukhatvdhga<m>. The retinue is absent 
only in the sense that the deities are not positioned around Heruka and Vajravarahi. 
Instead the twenty-four Yogim-Vira couples are installed from the head of Heruka 
down to his feet, and the four Yoginis of the innermost circuit and the eight of the 
outermost are installed in the twelve objects in his hands. 

413 Abhidhdnottara B f. 50v5-6: tatpardvrttyd sadvajram vajrasattvam vibhdvayet | 
trimukham sadbhujam caiva trinetram karundrasam | . . . (ff. 52v5— 53r3) anena 
codito ndtho bijam utpannam uttamam | kuhkumdkdravarndbham vajracihna- 
samutthitam | *sanmukham (corr. : khanmukham Cod.) dvadasabhujam 
vdrdhydsamalamkrtam | *sada(?)viramahdviram ardhaparyahkasamsthitam | tri- 
netram hasitam raudram kardlam bibhatsam Helihdnanam (em. : lelihdnalam 
Cod.) karundrasam \ bhairavam kdlardtrim ca pdddkrdntatale sthitam | 
athavdlidhasamsthdnakrtayogam *mahddbhutam (conj. : mahadbhutam Cod.) | 
. . . (f. 53r5-v2) *brahmanah (em. : brdhmana Cod.) krttim utkrttya prsthaprdvrta- 
vigraham | raudrabhairavacarmena *katim (corr. : katir Cod.) dvestya samsthitam | 
kapdlakhatvdhgadhara<m> asi-utpalasaradhdrinam \ ahkusapdsadamarumunda- 
cdpadharam tathd | tadvaktrdyudhavdrdhyd mahdrdgapade sthitd | jahghddvaya- 
samdslistd mahdsurata*sundarl (corr. : sumdhard Cod.) | mundasragddmadehogrd 
sanmudrdcihnabhusitd | evam bhdvayate yogi manjuvajramahdsukham. 


Genesis and Development of Tantrism 

armed, 414 and in the eleventh, where a six-faced, twelve-armed Heruka wearing 
the flayed skin of Rudra on his back embraces a twelve-armed Vajravarahi; 415 
and in the twentieth, in which a red five-faced and twelve-armed Heruka em- 
braces a Vajravarahi with same colour and hand-attributes. 416 

The literature also teaches Kalpas in which Vajravarahi is worshipped in her 
own right in the centre of a circuit or circuits of Yoginis. She may be one-faced 
and two-armed, as when she is worshipped as Heruka's consort, standing in the 
warrior pose at the centre of the circle of the eight cremation grounds, naked, 
red and menstruating, her face contorted with anger, with large fangs, three 
red eyes, wearing a chaplet of five skulls framed by two rows of Vajras, with 
crossed Vajras on her unbound hair, wearing a garland of fifty heads, which are 
not desiccated, as they are when she is Heruka's consort, but, like his, freshly 
severed and dripping blood. She holds aloft a red Vajra in her left hand with 
her index extended, a skull-bowl full of blood in her right, and a long white skull- 
staff resting in the crook of her left arm, She may possess, as before, only the first 
five of the six Mudras; but some emphasized her pre-eminence by requiring that 
since she is now the central deity of the Mandala she should also be smeared with 
ashes. She is surrounded by the thirty-six Yoginis, disposed as in the Mandala 
of Heruka, but with the difference that the Yoginis, like her, wear garlands of 
freshly severed heads, 417 or by only the inner circuit of four, or with no retinue 

414 Abhidhanottara B ff. 71r3-72v5: vajrasattvapardvrttyd herukatvam vibhdva- 
yet | pancdnanam dasabhujam vdrdhydsamalamkrtam . . . (f. 72v4-5) tadvarna- 
bhuja* samsthdnd (corr. : samsthdnam Cod.) muktakesi tu nagnika vydghracarma- 
nivasand khandamanditamekhald \ kapdlamdlini raudri karundrdgasuvihvald . 

415 Abhidhanottara B ff. 79v3-80r6: sadvaktram viram bibhatsam srhgdrahasitam 
raudram lelihdnanam \ sanmudrdmudritam deham nanabharanamanditam \ 
vdrdhyd *tu samdpannam (em. : nusamdpannd Cod.) jdnudvayasuvestitam . . . (f. 
80r2) rudracarmdmbaradharam . . . (f. 80r5-6) tadvarnabhujasamsthdnd mukta- 
kesi tu nagnika. 

416 Abhidhanottara B f. 113r3— v4: herukdkdram dtmdnam ddkinicayapardvrtam \ 
mahogram raktavapusam pancajnanodbhavodbhavam | raktam nilam ca haritam 
pitam santasitordhvakam | trinetram dvadasabhujam alidhapadasamsthitam | 
. . . (f. 113v3-4) agrato vajravarahya tadvarnayudhadharini. 

417 This is the main Kalpa taught in the Abhisamayamahjari (pp. 131, 1. 9-133, 1. 
1). I propose the following emendations and corrections to the text of the pub- 
lished edition: for mithya drstiprahana vikrtaikananam (p. 131, 1. 15) read mithya- 
drstiprahanad vikrtaikananam ; for cakrikundalakanthikdrucakakhatvahga- 
mekhalakhyapancamudradharam (p. 131, 1. 18) read cakrikundalakanthikarucaka- 
khandahkamekhalakhyapancamudrddharam; for iti kecit | mandalanayikatvena 
sanmudritam ity eke read iti mandalanayikatvena sanmudritam ity eke (p. 
132, 1. 3); for vajravalldvayamadhyakrta- read vajravalidvayamadhylkrta- (p. 
132, 1. 9); and for astavijhanam nairatmydsvarupatvena read astavijndndndm 
nairdtmyasvarupatvena (p. 132, 1. 12). 


The Saiva Age 

at all. 418 

There are other forms of this kind, among which one is particulary worthy of 
note because it shows her four-faced and twelve-armed like Heruka himself, his 
equal as it were or, rather, the fusion of both within her, since her fanged face is 
divided down the middle into a male half on her right and a female half on her 
left (ardhanarisvaramukha), a Sakta reflex of the well-known Ardhanarisvara 
image of Siva. She has the same hand-attributes as the twelve-armed Heruka 
except that the battle-axe and trident have gone, an elephant-goad has taken the 
latter's place. The hand that held the skull-staff now holds the skull-bowl, the 
skull-staff rests in the crook of that arm, and the two hands that are now free 
form the flame gesture (jvalamudra) on her forehead. The place of the elephant 
hide is taken by the flayed skin of a man. She holds the Vajra and bell in her 
crossed principal hands and turns them over each other in the gesture known as 
the revolving lotus (kamalavartah). She is red, naked, and intoxicated with pas- 
sion, adorned with all six Mudras, the new moon and crossed Vajras on her hair, 
a chaplet of skulls above her forehead, and the bone-filigree around her hips. 
She dances wildly in the centre of her retinue, visualized at the moment that she 
stands with her left leg on the ground flexed at the knee and her right foot raised 
and placed on the inside of her left thigh with the right knee turned out. She 
is surrounded by the thirty-six Yoginls with the addition of the four goddesses 
Mamaki, Locana, Tara, and Pandaravasin of the Guhyasamdja Yogottara sys- 
tem. The four innermost goddesses have the heads of a lion, sow, elephant, and 
horse, and hold in their four hands the skull-bowl, skull-staff, head of Brahma, 
and chopping-knife. Outside them are the four Yogottara goddesses, each at the 
centre of a lotus with six petals, six-armed and adorned with the six Mudras. 
They hold in one of their two principal hands the symbol of the Tathagata-family 
to which each belongs (a Vajra, a wheel, two crossed Vajras, and a lotus respec- 
tively) and in the other a bell, turning them over each other. In the other hands 
they hold a skull-bowl, the head of Brahma, and a rattle-drum, with a skull-staff 
in the crook of the principal left arm. The twenty-four Yoginls of the sacred sites 
are placed in groups of six on the petals of these lotuses. They are four-armed, 
and hold the symbol of the Tathagata-family of the Yogottara goddess on whose 
lotus they are placed, a skull-bowl, a skull-staff, and a rattle-drum. They wear 
chaplets of skulls and show only five of the six seals. Like the central goddess 
they are half male and half female {ardhanarisvaryah). All the goddesses in the 
Mandala up to this point are naked and dancing. Outside them is the final circuit 
of eight Yoginls. The four in the four doors of the Mandala, with the heads of a 

418 Abhisamayamanjan, p. 142, 11. 13—19. 


Genesis and Development of Tantrism 

crow, owl, dog, and sow, stand naked in the warrior-pose, dwarfish, with squint- 
ing eyes. The four in the corners have the heads of a buffalo, an ass, a camel, and 
a horse, and like all but the door-guardians, are visualized in the dance posture. 
All eight of these outer Yoginis have the five Mudras and chaplets of skulls, and 
are four-armed, holding a skull-bowl, the head of Brahma in their left hands, and 
a chopping-knife and rattle-drum in their right. 419 

The cult of the independent goddess (BhagavatI) appears to have been 
a particularly vigorous development, to judge from the exceptionally large 
number of variant forms that emerged. 420 Within the earlier scriptural lit- 
erature the Abhidhdnottara contains several sections devoted to Sadhanas of 
Vajravarahi; 421 in the Herukdbhyudaya eleven of its forty-four chapters are 
devoted to her Mantras and their procedures; 422 and the section of the Tenjur 
devoted to the Cakrasamvara cycle (Toh. 1403-1606) contains over sixty texts 
devoted to the varieties of her cult as Vajravarahi or Vajrayogini (Toh. 1541- 
1606). Sakyaraksita, a pupil of Abhayakaragupta (1064-1125), after detailing 
the Sadhana of several of her forms in his AbhisamayamafijarT, 423 adds that 
these are but a few of the many that were current in his time: 424 

So it should be understood that in accordance with the various mentalities of 
those requiring to be trained there are countless traditions of the Goddess such 
as this, transmitted through the generations from teacher to pupil in accordance 
with the [founding] instruction of various Siddhas. What I have shown here is no 
more than an indicative fraction of the whole. 

This Sakta trend is also evidenced in the practice of the Newars of the Kath- 
mandu valley down modern times. For their ceremony of initiation before the 
Mandala of Cakrasamvara is followed on the final day by initiation before 

419 This form is taught in Abhidhdnottara ff. 63vl-70r4 (Patala 9 in the enumera- 
tion of this manuscript), from which it entered the Varahyabhyudaya. A lightly 
adjusted version of this Kalpa is found in the collection of Sadhanas of Va- 
jravarahi/Vajrayogim that came to bear the title Guhyasamayasadhanamala in the 
colophons of later manuscripts; see English 2002, pp. 54-59. 

420 See English 2002 for an illustrated survey of these variants. 

421 Patala 12/9: Varahi Vajrayogini (4-faced, 12-armed; ardhandrisvarimukha); 22/19: 
Mrtasamjivam (4-faced or 8-faced, 16-armed); 36/33: Vajravarahi (3-faced and 
6-armed or 6-faced and 12-armed, surrounded by Guhyottama etc.); 37/34: Va- 
jravarahi surrounded by Yamini etc. 

422 Patalas 6, 8-11, 23-24, 29-31, and 34. 

423 rp^g AbJii sama y ama fij ar i i s ascribed to Subhakaragupta in its sole edition. This is 
an error and goes against the evidence of the colophons of the manuscripts (EN- 
GLISH 2002, p. 357, n. 6). 

424 Abhisamayamahjari, p. 152: tad *evamadayah (em. : evam adaya Ed.) siddho- 
padesaparamparayata vineyasayabhedad ananta bhagavatya amnaya boddhavyah 
| dihmatram idam darsitam. 


The Saiva Age 

the Mandala of Vajradevi (Vajravarahi). 425 Nor was this confined to the sub- 
continent. In Tibet too Vajravarahi/Vajrayogini rose to a position of special 
honour, notably in the bKa' brgyud and Sa skya traditions, but also in later 
times among the dGe lug pas, rNying ma pas, and Bon pos. 426 

There are other compilations, scriptural and secondary, that survive in 
Nepalese manuscripts but did not reach Tibet, which attest her prominence 
in the last phase of the Mantranaya: the Vajravarahikalpa, of about three 
thousand verses, which interweaves the Dakarnava and the Samvarodaya, 
and incorporates thirteen non-scriptural Sadhana texts of Vajravarahi and 
one of Nairatmya, the consort of Hevajra; 427 the closely related Yoginijala, 
of about one thousand verses; and the collection of forty-six Sadhanas of 
Vajrayogini known as the Guhyasamayasadhanamala. 428 Moreover, two texts 
devoted to the cult of this goddess were added to the canon of scripture re- 
ceived by the Tibetans. The first is the Vdrdhyabhyudayatantra, a short 
work of three hundred verses counted among the explanatory Tantras of 
the Laghusamvara but consisting almost entirely of passages lifted from the 
Samputodbhava, the Abhidhdnottara, and the Samvarodaya; 42 ® and the second 

Gellner 1992, pp. 273-279. His account of the ceremonies is based upon what he 
was told by the late Asha Kaji Vajracharya (ibid., p. 273). That the Cakrasamvara 
initiation is followed by a separate Vajradevi initiation is confirmed by the evidence 
of the Diksavidhi, the manual in the Newari language that guides these rituals. 

; See English 2002, pp. xxii-xxvii. 

I have not yet undertaken a thorough analysis of the whole text. The interweav- 
ing that I report is of Dakarnava, Patala 2-3 and Samvarodaya 2-3 in the first 3 
Patalas. The nidanavakyam of the Samvarodaya is borrowed with the substitution 
of vardhibhagesu for the Samvarodaya's yoginibhagesu. I have noted the incorpo- 
ration of the following Sadhana texts (identified here with the numbers ascribed 
in BHATTACHARYA's composite Sadhanamala): 217-218 in Patala 36, 219-225 in 
Patala 37, 226-228 and 231 in Patala 38. 

This is the title under which the work has been catalogued in Tsukamoto et al. 
1989, p. 285. It is based, I surmise, on the colophon of the last Sadhana in the 
collection, the Ddkiniguhyasamayasddhana of Anangayogin. 

1 The correspondences are as follows (S = Samputodbhava; LS = Laghusamvara; AU 
= Abhidhdnottara; SU = Samvarodaya): 1.5-6b = S 6.3.26-27b; 1.17 = S 6.3.44c- 
45b; 1.18ab = S 6.3.45cd; 1.20cd ~ S 6.3.46cd; 1.21 = S 6.3.47; 1.31 ~ S 6.4.39; 
1.33-43b = S 6.4.40-50; 2.15 = LS 1.19; 2.17c-18 = S 6.3.2-3b; 2.24-27b = S 6.3.3c- 
6; 2.27cd = S 6.2.2ab and 6.3.7ab; 2.28-29 = S 6.2.2c-4b; 2.31-33b = S 6.2.4c-6b; 
2.34-40 = S 6.2.6c-14; 2.43-44d = S 6.2.15c-16; 3.1-2 = S 6.2.27-28; 5.8-14 = 
S 6.3.11-17; 6.1-2 = SU 7.1-2; 6.3b-6b = SU 7.14c-17; 6.6c-12b = S 6.3.35-40b; 
6.14-19b = S 6.3.40c-45; 6.23-30 = AU 14.58-65; 7.3-7 = S 6.3.19c-24; 8.3-5 = AU 
3.8c-llb; 8.17c-18 = AU 16.2-3b; 8.20b ~ AU 16.3c; 8.20c = AU 16.4a; 8.21-22 * 
AU 16.4b-5; 8.24-37 = AU 16.6-19; 8.39-41 = AU 16.23-25; 9.1c-5 = AU 4.3-7b; 
9.6-17a = AU 4.9-20b; 9.21-39a = AU 4.24-38f; 9.39c-41b = AU 4.42c-44b; 9.41c- 
44 = AU 4.39-42b; 9.45-51 ('47', '48' and '50' are Mantras) = AU 4.44c-46 (with the 
same Mantras); 9.52ab = AU 4.51ab; 9.54ab = AU 4.51cd; 10 = AU 50. 


Genesis and Development of Tantrism 

is the Vidyadharlkramavajrayoginisadhana, which appears in the Kanjur (Toh. 
380) between the major Tantras of the Cakrasamvara cycle and those of con- 
tested authenticity, 430 included perhaps, in spite of its genre, because it states 
in its opening words that it is part of the otherwise unattested Mahamayajalo- 
rdhvajatottaratantra, which, it claims, was extracted from the Trilaksa, that it 
to say, from the vast mythical Ur-text of this cycle, the Trilaksabhidhana. 431 

Further evidence of this Sakta trend is seen in the views of the tradition 
concerning the nature of the revelation of this Ur-text, which, it was claimed, 
contained the required Buddhist preamble (nidanavakyam) that is lacking in 
the Laghusamvara itself. Bhavabhatta, taking care not to claim direct access to 
that mythical source, saying only that his knowledge of its nidanavakyam has 
reached him through the lineage of his teachers (guruparampara), 432 asserts 
that it reveals that the teacher of the Tantra was Bhagavan Mahavajradhara, 
the requester his consort Bhagavati Vajravarahi, and the reciter Vajrapani. 
These then, it follows for Bhavabhatta, are the dramatis personae of the 
Laghusamvara too. But he reports a contrary view that Vajravarahi was the 
teacher and Mahavajradhara her pupil. 433 The imposition on the text of the 
claim that it is a dialogue between the deity and his goddess-consort brings it 
into line with the Saiva scriptural literature of the Vidyapitha. For there the 
Tantras take the form of Bhairava's teachings in answer to the questions of 
the Goddess (Devi/Bhairavi). In the explanatory Tantras of the Cakrasamvara 
cycle this model is made explicit in the Vajradaka, where Vajrasattva/Vajradaka 
teaches in response to the questions of Devi, and in the Dakarnava and Va- 
jravarahikalpa, where Viresvara responds to the questions of Vlresvari. But 
in the Caturyoginisamputa, another of the satellite Tantras of this cycle, the 
goddess Vajrini (Vajravarahi) is the teacher and Vajrin (Heruka) the ques- 
tioner. 434 That this inversion seen in the view reported by Bhavabhatta and 

430 In Sanskrit it is preserved as the twenty-first Sadhana in the Guhyasamaya- 
sadhanamala , ff. 85r4— 86rl. 

431 Guhyasamayasadhanamala, f. 62r2: athatah sampravaksyami trilaksakrstamaha- 

432 Bhavabhatta, Cakrasamvarapahjika, introduction: mahavajradharo desakah. 
. . . bhagavati vajravarahy adhyesika vajrapanih samgata . . . vajravarahy- 
adhyesitasya bhagavatah prativacanam etad athata ityadi . . . adhyesika deviti ko 
niyama iti cet | guruparamparato hi sruyate mulatantre saivadhyesiketi | tata ihapi 
saiveti gamy ate. 

433 Ibid., following the preceding citation: bhagavan adhyesako bhagavati desiketi 
kecit. acintyarupo hi tathagatanam abhiprayah 'Some say that the Lord 
[Mahavajradhara] was the requester and the Goddess [Vajravarahi] the teacher. 
For the intention of the Tathagatas is inscrutable'. 

434 Caturyoginisamputa 2.15d-16: atha sa vajrini devi idam vakyam udirayet | 
abhisekam *sukathitam (conj. [=legs par brjod nas Tib.] : kathitam Cod.) 


The Saiva Age 

in the Caturyoginlsamputa is evidence of a more Sakta tendency within the 
tradition is obvious in itself, but it is confirmed by parallel practice in the 
most Sakta of the Saiva scriptures, namely the Kalikulakramasadbhava, the 
Kalikulapancasataka, and the Manthanabhairava. 

The Adoption of the VidyapItha's Carya and Yoga. As for the 
practice of initiates into this tradition, that too shows increased saktization. 
For it now enacts the iconography of their deities through the adoption of 
the VidyapItha's Kapalika mode of post-initiatory observance (carydvratam). 
Buddhist Sadhakas now carry the skull-bowl (kapdlam) and skull-staff 
(khatvangah), and put on the Mudras of human bone and a brahmanical thread 
(yajnopavitam) made of the twisted hair of corpses or human sinew, and dust 
their bodies with ash. 435 

*ganamandalam eva ca (conj. [=tshogs kyi dkyil 'khor nyid dag dang Tib.] : lack- 
ing in Cod.) | aparam kathayisydmi devatdnydsam uttamam 'Then that goddess 
VajrinI uttered the following words: I have fully explained the initiation rites and 
the Ganamandala. Next I shall explain the supreme [rite of the] installation of the 
deities'. For the verb udirayet as a past indicative cf. Pali udlrayi. 
435 E.g. Yogaratnamala on Hevajra, p. 155: carydkdle ganacakrddau va pancdndm 
mudranam dharana; Laghusamvara f. 37v3 (51.2): nivasanam pahcamudrddi 
gdtrasya; Abhidhdnottara B f. 10v2-2 (3.18): pahcamudrddharo nityam kapala- 
krtasekharah \ kapdlakhatvdhgadhdri ca bhasmoddhulitavigrahah; Bhavabhatta, 
Cakrasamvarapanjika on Laghusamvara 51.21a: pancamudrdditi. kanthikdcuda- 
keyurakundalabrahmasutraniti; Jayabhadra, Cakrasamvarapanjika on Laghu- 
samvara: p. 128: panca mudra rucakasiromanikundalakanthikayajnopavitah 
panca | sarvada tair avirahito bhavet; Yoginisamcara 6.12c-13d: kanthikarucaka- 
kundalasiromanivibhusitah yajnopavitam bhasmeti mudrasatkam prakirtitam; 
Khrag 'thung mngon par 'byung ba f. 13r4 (Herukabhyudaya 15.27): nub mo ru 
ni dam tshig ste | dpa' bo rtag tu gcer bu yin | sgrub pos sngags dang phyag 
rgya dang | phyag rgya Inga dang yang dag Idan 'Observing the vows (samayl), 
the Sadhaka Hero (vlrah) [should] always [be] naked at night (ratrau ca satatam 
nagnah [?]), equipped with the Mantras and Mudras (mantramudranvitah), 
and wearing the five [bone] Mudras (pancamudrasamanvitah)'; Hevajra 1.3.14: 
cakri kundala kanthi ca haste riicaka mekhala \ pancabuddhavisuddhya ca eta 
mudrah prakirtitah; 1.6.2a: sirasi cakri dhartavya (= siromanih, a circlet of 
bone; the mekhala is a filigree made of small pieces of bone worn around 
the hips); Hevajra 1.6.16cd: bhasma kesapavitram ca yogi bibharti caryaya; 
Muktavali ad loc: kesapavitram kesayajhopavitam; Vajravali B, p. 218: athava 
nrnaharumayam kesakrtam va brahmasutram 'or the sacred thread may be 
made of human sinew or hair'; Abhisamayamahjari, pp. 131—132: cakrikundala- 
kanthikdrucakakhanddhkamekhaldkhyapancamudrddharam (see here p. 174) | 
kanthikdrucakakundaldni siromanivibhusitam \ yajnopavitam bhasmeti mudra- 
satkam prakirtitam iti kecit. For the Saiva case see, e.g., Svacchandoddyota 
on 3.2b: mudrdlahkdrabhusitah sikhdkarnaprakosthapratisthdpitapancamudrah; 
Picumata, f. 101r3 (21.104): karnau sirasi bdhubhydm asthikhandair vibhusitah; 
a verse cited by Yamunacarya in his Agamaprdmdnya, p. 93 (Y), edited here by 
collation with the closely related verse cited by Nirmalamani as cited by Brun- 
ner in Somasambhupaddhati vol. 3, p. 681, n. 7 (N): *kanthikd (em. : karnikd 


Genesis and Development of Tantrism 

The pan-Indian topography of the Sakta Saivas' sacred sites, their Pithas, 
Ksetras, Upaksetras, Samdohas/Chandohas, 436 and the like, is also adopted. 
Two lists of such sites are found: one in the Vajradaka and the other in the 
Laghusamvara} 31 Also adopted is the practice of visiting these sacred sites 
(pithabhramanam) 438 in search of meetings with the Yoginis/Dakinis that are 

Y : kundikd N) kundalam caiva *rucakam (Y : uragam N) ca *sikhdmanih (nih 
N : nim Y) | *bhasma yajnopavltam ca (Y : kesayajhopavltam ca N) *mudrd- 
satkam pracaksate (Y : mudrd ete mahdvratdh [< mahdvratej) 'The [Kapalikas] 
teach that the six Mudras are (1) the necklace, (2) the earrings, (3) the bracelets, 
(4) the hair-jewel, (5) ashes and (6) the sacred thread [made from human hair]'. 
This followed in Y by a second verse: kapdlam atha khatvahgam upamudre 
prakirtite | dbhir mudritadehas tu na bhuya iha jdyate 'The skull-bowl and skull- 
staff are called the sub-Mudras. One whose body is sealed by these [eightl is not 
born again in this [world]'; Jayadrathaydmala, Satka 3, f. 201v3: dvitlyam tu 
vratam vaksye ghorakdpdlarupina<m> \ sire kapdlamukutam siromdldvibhusitam \ 
kare karnau tathd pddau asthikhandair vibhusitau \ vdme kapdlam khatvahgam 
tathd vai daksine kare. The six Mudras minus the ashes, that is to say, the 
five of the Buddhist lists, are defined, but not numbered, in Jayadrathaydmala, 
Satka 1, f. 139rl-3 (23.33-36b), in the order earrings, bracelets, hair-jewel, sa- 
cred thread of human hair, and necklace: virdndm nrpasdrdula tantre 'smin 
bhairavdrcite \ subhrasahkhe prakartavye dvyahgule karnike subhe \ *rucake (em. : 
caruke Cod.) dvyahgule saste turydhgusthah sikhdmanih | trivrnnarakacotpannas 
tripahcasarikah samah | kanthdj jaghanasamsparsT (*ja corr. : jah Cod.) sastah 
pahcavato 'pi ca || suvrttamanisamghd*ta(corr. : tah Cod.)samghdtaikdvali samd | 
dhdryd sddhakacandrena sesabhutd tadiccha*yd (em. : gd Cod). The 80th chap- 
ter of the Picumata describes, but does not number, (1) the hair-jewel, (2) earrings, 
(3) a necklace (kanthamudrd), (4) the sacred thread, and (5) ornaments of bone 
on hands, arms and hips. The last takes the place of the bracelets (rucake) listed 
elsewhere and in Vajrayanist texts (Picumata ff. 311v-312r): cuddmanikapdlena 
sikhdydm yo nivesitah \ isvaras tatra vijheyo adhidevo vardnane \ jhdnasaktih 
kriydkhyd ca karnike parikirtite \ kanthe sthitd tu yd mudrd aham tatrddhidevatam 
| rudro mdtrganaih sdrdham jhdtavyas tu vardnane | anantd hy upavite tu saktih 
sarvddhvagd para | hastabdhukatisthais ca visnur jheyo 'dhidevatam \ saktayo 
vividhdkdrd jatdndm adhidevatam \ etan mahdrthadam deviyo vijdndti tattvatah \ 
sivavat sa tu boddhavyo viruddhdcarano 'pi yah. 

436 rp^g g a j va t erm samdohah for one class of site consistently appears in Buddhist 
treatments in the form chandohah (e.g. Laghusamvara 50.22 and Hevajra 1.6.10). 
This substitution of initial ch- for s-/s- is probably an east-Indianism; cf. Oriya 
chahcibd < Skt. samcayati; Bengali chdtu < Skt. saktuh; Oriya chdc, chacd < Skt. 
satya-; Bengali chut, Bengali and Oriya chutd < Skt. sutram; Oriya chana < Skt. 
sanah; Bengali chddld < Skt. sddvalam; and Bengali chikal, chikli < Skt. srhkhala-, 

437 On these lists see here pp. 192-203. 

438 See, e.g. Samvarodaya 8.29b,d: pithddidesagamanena visuddhadeham . . . vande 
sadd guruvaram sirasd natena 'At all times, with head bowed, I venerate the 
best of Gurus, . . . whose body has been purified by going to the Pithas and other 
[such] sites'; 9.25: pithopapithasevandn nirmalo bhavati mdnavah \ bhraman 
nimittam samlaksya nirvikalpena dhlmatah 'A man becomes pure by frequent- 
ing Pithas and Upapithas. The adept should wander [there] without hesita- 
tion, observing [any] signs [that may arise] without inhibition'; 26.14 . . . 18c-19: 


The Saiva Age 

believed to frequent them and to be incarnate there in human women enlight- 
ened from birth or in childhood; 439 classifying such women as belonging to one 

pithe ksetre ca cchandohe meldpakasmasdnake || pujyapujakasambandhe amrtam 
argham uttamam || . . . pratisthdhomakdlesu pithabhramanagocare || naimitte yo- 
ginlpujye mantrasddhanatatksane | evam bahuvidhd jneyd tasya doso na vidyate 
'In a Pitha, Ksetra, Chandoha, Melapaka, a cremation-ground, or an encounter 
between worshipper and worshipped, wine is the highest guest-water. ... on the 
occasion of installation ceremonies, when wandering through the Pithas, during 
worship of the Yoginis occasioned by some event, and when doing the Sadhana of 
a Mantra. He should know that there are a manifold [occasions] such as these [on 
which he may drink wine]. He will not be at fault'. Cf Nisisamcdra, f. 10v2-3: 
evam eva prakdrena ghorasddhanatatparam | ksetra paryatamdnasya sddhakasya 
mahddhiye | sabdam dadati yah kascit tasya prasnam vaddmy aham 'O you of great 
understanding, I shall teach [you] the requests [that should be addressed] to any [di- 
vine being] who speaks to the Sadhaka as he wanders in this manner visiting the 
Ksetras, intent on the Ghorasadhana'; Tantrdloka 29.40ab: iti samketdbhijho bhra- 
mate pithesu yadi sa siddhlpsuh 'If a person seeking Siddhis wanders from Pitha to 

Pitha knowing these signsf, the chummdh] 

439 Jayabhadra, Cakrasamvarapanjika on 26.1, p. 125: ydvanti ksetropaksetrdni yo- 
gapithdni tatra vyavasthitd dutyah siddhidds cumbandvaguhandd etdh viseseneti 
ydvat 'Dutis are present in all the Yogapithas, the Ksetras, and Upaksetras. 
These bestow Siddhi, especially through kissing and copulating [with the 
Sadhaka]'; Laghusamvara 41.4c-5, reconstucted from the lemmata in the 
Cakrasamvarapanjika of Bhavabhatta, the commentary Sddhananidhi of Kam- 
balapada (K), this passage as incorporated in Vajraddka f. 41v2 (18.2) (V), and the 
Tibetan translation (T): sarvottaresu pithddi ddkinyas tu sarvavydpini \ dese dese 
*'bhijdyante (V, mngon par skye T : jdyante K) jndnayuktdh svayonisu \ ddkinyas 
tdh samdkhydtdh vajramandalandyikdh 'In all these superior [sites] in various re- 
gions, namely the Pithas and the rest, women are born who are endowed with 
knowledge in their mother's wombs. It is these that are called Dakinis, leaders 
of the Vajramandala'. Cf. Tantrasadbhdva f 115v3-4 (16.279c-280): vijhdna-m- 
udayamm dsdm kathyamdnam nibodha me | plthajds cdstabhir varsaih ksetrajd 
dvddasdbdikdh | dvdre sodasabhir devi yonijdh saptavimsati 'Listen to my account 
of the emergence of the enlightenment of these [Yoginis]. Those born in Pithas 
[achieve it] at the age of eight, those born in Ksetras at the age of twelve, [those 
born in] Dvaras at the age of sixteen, and those born of [lowly] wombs at the 
age of twenty-seven'. Cf. Tantrdloka 15.97cd-100b: bdhye tu tddrsdntahsthayoga- 
mdrgavisdraddh || devyah svabhdvdj jdyante pitham tad bdhyam ucyate | yathd 
svabhdvato mlecchd adharmapathavartinah || tatra dese niyatyettham jhdnayogau 
sihitau kvacit | yathd cdtanmayo 'py eti pdpitdm taih samdgamdt || tathd pithas- 
thito 'py eti jndnayogddipdtratdm 'In the outer [Pithas, Ksetras and the rest as 
opposed to these transposed into the person of the worshipper] divine women are 
born who are innately adept in the path of such internal meditation. Just as the 
barbarians of other lands naturally follow paths outside of ordained religion, so in 
some [women] in these places enlightenment and meditation-trance are naturally 
present. And just as a person becomes a sinner through association with those [bar- 
barians], even though he makes no effort to assimilate, so a person residing in a 
Pitha becomes the beneficiary of enlightenment, meditation-trance, and [Siddhis]'; 
and 29.40: iti samketdbhijno bhramate pithesu yadi sa siddhipsuh \ acirdl labhate 
tat tat prdpyam yad yoginivadandt 'If a person seeking Siddhis wanders from Pitha 
to Pitha knowing these signs[, the chummdh], he quickly attains from the mouths 


Genesis and Development of Tantrism 

or other a fixed number of deity -clans (kulam) and of specifying various charac- 
teristics of appearance and behaviour that enable the adept to determine these 
clan-affiliations; 440 the consumption and offering of meat and alcoholic liquor in 
their rites; 441 the consumption of foul substances without inhibition as an ini- 
tiatory test of nondual awareness; 442 the sacrifice and consumption of the flesh 

of Yoginis whatever he wishes'. 

440 Laghusamvara, Patalas 16-24 (> Abhidhanottara, Samputodbhava, Samvarodaya, 
Mahdmudrdtilaka, Vajraddka); and parallel passages in the Vidyapitha texts Yo- 
ginlsamcdra , Tantrasadbhdva, Siddhayogesvarimata, and Picumata. For full refer- 
ences see Sanderson 2001, pp. 42-43 (Table I). 

441 Bhavabhatta, Cakrasamvarapahjikd, p. 497: dsu pujanlyd madyais ca mdmsair 
api vajradevyah | tdh pujitd bhaktimato janasya srlherukasydbhiratim gatasya 
samtustacittd varadd bhavanti 'On these [lunar days] [the women who embodyl 
the Vajra goddesses should be worshipped with offerings of alcohol and flesh. When 
they have been worshipped they become delighted and bestow boons on any devotee 
who is attached to Heruka'; Abhidhanottara B f. 48v5- (6.50d-56a): vividhai<h> 
samayottamaih || *madyair (em. : padma Cod.) ndndvidhai<s> caiva surdpdnais 
tathottamaih | *virameldpakam (vlra corr. : vird Cod.) divyam yogini vivi- 
dhottamd<h> || kapdlakhatvdhgakard<h> kartikddamarukottamd<h> | vddyai<r> 
ndndvidhair divyai<r> bhojyabhaksyarasottamaih || vividhais cumbandlihgais 
cosyalehyottamottamaih \ evamvidham smasdnam tu yaksavetddardksasaih || 
balim tatraiva ddtavyam *herukarupam (em. : heruko rupam Cod.) udvahet | 
damaruvajraghantd<m> ca vddyanrtya<m> prakurvati || digvdsd mudrayd yukto 
humphatkilakildyate | dlidhapadayogena jvdldmudrdm tu bhdvayet \\ mukham 
dpurya samayaih 'The illustrious assembly of Viras [with Yoginis should be cele- 
brated! with [the eating of] the various superior sacramental meats [detailed above], 
with various wines and excellent draughts of rice-beer. The various Yoginis, holding 
the skull-bowl, skull-staff, a chopping-knife, and a rattle-drum [should be gratified] 
with various forms of music, the savours of excellent foods soft and hard, with kisses 
and embraces, with foods to be sucked and licked. Such [should be] the cremation 
ground [on this occasion]. There he should offer Bali to the Yaksas, Vetalas, and 
Raksasas. He should assume the form of Heruka. He should [sound] the rattle- 
drum and Vajra-bell, dance, and make music and dance. Naked together with his 
consort (mudrd) he utters the syllables HUM PHAT and cries of joy. Standing in 
the warrior pose he should make the Flame Mudra with his hands, having filled his 
mouth with the sacramental meats'. Patala 16 of the Samvarodaya is devoted to the 
preparation and use of alcoholic drinks. At its end (16.51abc) it says: madyapdnam 
vind pujd homas caiva ghrtam vind | sadgurum ca vind dharmam 'There cannot be 
worship without drinking wine, fire-sacrifice without clarified butter, or religious 
practice without the Guru'. Cf. the scriptural passages on the indispensability of 
wine in Kaula worship cited by Jayaratha on Tantrdloka 29.1-13. One of those 
passages says that beer is the Goddess and wine Bhairava; surd ca paramd saktir 
madyam bhairava ucyate (p. 9, line 2). Cf. Samvarodaya 16.12cd: yd surd *vajrayo- 
ginyd (conj. : vajrayoginyo Ed.) yo madah sa ca herukah 'Beer is Vajrayogim and 
wine is Heruka.' 

442 See, e.g., Kumaracandra, Herukdbhyudayapanjikd, p. 156: tatreti mandale 
'mbhojabhdjane samskrta<m> biddlavidddikam daksindbhimukhdcdryo vdso- 
baddhdsyam sisyam dnlya omkdrddigdyatryd raksitvd *potahglpratipotahgl- 
prasnottarakriydpurvakam (corr. : potanglm pratipotahgim Ed.) pravesya taddsye 
nivesayet 'There, that is to say, before the Mandala, the Acarya, facing south, 


The Saiva Age 

of human beings believed to have been reincarnated seven times for this pur- 
pose (saptavartah), recognized in both traditions on the basis of similar physical 
characteristics, and the use of their skulls as skull-bowls; 443 the practice of visu- 
alizations in which the Sadhaka enters the body of a victim through the channels 
of his vital energy (nadi), extracts his vital essences, and draws them into him- 
self; 444 that of yogically raising one's consciousness out of one's body through 

should sacramentalize in a skull-bowl some substance such as cat excrement. He 
should then lead the blindfolded candidate forward, protect him with the Gayatri 
[of Heruka] beginning with OM, and after addressing him with the word POTANGlf, 
the chommd of welcome] and having received [the chommd] PRATIPOTANGI in re- 
sponse, he should bring him before [the Mandala] and place that substance in his 
mouth'. For the Saiva literature see the passages cited in SANDERSON 2005c, pp. 
113-114, fn. 63. 

443 See, e.g, Laghusamvara f. 10r3-4 (11.1-2) and 49.4-13 (49.4-8 = f. 35v5-7; 49.8-13 
= bDe mchog nyung ngu, f. 244r2— 5); Abhidhanottara, Patala 63; Herukdbhyudaya, 
Patala 13 (Khrag 'thung mngon par 'byung ba f. 10r7— v6); Hevajratantra 1.11.10— 
11; Mahdmudrdtilaka f. 23r3-4 (12.20-21): tddrsam yatndt saptajanmdnam dnayet 
| ndndpujopahdrena pujayet tarn samdhitah || tasyottamdhgam utkrtya kdrayet 
padmabhdjanam | tatraiva pdtre madanam pdyayet prajhayd saha 'He should 
with all effort bring such a man of seven rebirths. With concentrated mind 
he should honour him with the various offering-substances. Having decapitated 
him he should make the head into a skull-bowl. In that vessel he should drink 
wine with his consort'; f. 51r5-v2 (24.1-3c): athdnyam *caiva (conj. : caika Cod.) 
karmdkhyam pravaksydmy ddardc chrnu \yena prdsitamdtrena dsu siddhih pravar- 
tate || susnigdhas ca sugandhdhgah sugandhasvedamanditah | satyavddi salajjdtmd 
nivesati dram sadd | krpdparah ksdntiyutah satyavddi nirdsrayah | saptajanmd 
trijanmd vd. In the Vidyapitha literature see the treatments of this topic in 
Jayadrathaydmala Satka 3, Yoginlsamcdra , Kdlajhdnapatala; Tantrasadbhdva, 
Adhikdra 7; and Tantrdloka 16.63-64 and Jayaratha's introduction to this passage. 

444 See, e.g., Herukdbhyudayapahjikd on Herukdbhyudaya, Patala 13 p. 155: svadehdt 
ddkinlh sphdrayitvd sddhye gudena pravesya navadvdrair nddimdrgena pasoh 
sddhyasya *bijam (conj. : bijamjivam bijam Ed.) sukrddikam grdhayitvd niskdsya 
svadehe pravesayet 'He should emanate the Dakinis from inside his body, have 
them enter the victim through his anus [or any one of] the nine apertures and 
passing through the channels of the victim's vital energies, seize his seed, his se- 
men and other [vital essences]. Then he should have them exit [the victim] and 
return [with these] into [his own body]'; on Herukdbhyudaya, Patala 42, p. 167: 
athavd sddhyam dkrsya tacchukrddi pitvd bhaksayet 'Having attracted the vic- 
tim he should [extract and] drink his semen and other [essences], then eat [the 
flesh]'; Abhidhanottara B f. 51vl-3 (9.62-64b): vdrdhydtmabhdvena tarjanyd ndbhi 
vedhayet \ ddkinyddi tu cakrasthd devya<h> *sucydkrtis (em. : sucydkrtds Cod.) 
tathd || navadvdre *pravesyaitd (conj. : pravesya tarn Cod.) *vedhayed (corr. : vi- 
dhayed Cod.) dhrdayapahkajam \ yoginyd hata*mdtre (conj. : mdtram Cod.) tu 
pibet ksatajam uttamam || hatam ca bhaksayet so hi buddho bhavisyati ndnyathd 
'By identifying with Varahi he should pierce the navel [of the victim] with his in- 
dex finger [in the gesture of threat] and cause the Dakinis and other goddesses of 
the Mandala to take on the form of a needle [through visualization]. When he has 
made them enter [the victim in this form] through the nine apertures [of the body] 
he should have them pierce through the lotus of his heart. As soon as the Yoginis 
have killed him he should drink his excellent blood and eat his flesh. For it is certain 


Genesis and Development of Tantrism 

the central channel as a means of ending one's life and ascending to a paradise 
or liberation, a practice known as utkrantih in Saiva sources and thence in the 
Buddhist Yoginitantras (Tib. 'pho ba); 445 the adaptation of this practice as a 

that [thus] he will become a Buddha'; Mahdmdyd 2.10-14b. On the extraction of the 
vital essences by such yogic means in Vidyapitha sources see, e.g., Picumata f. lOvl- 
4 (3.198c-207): pravisya ca puram divyam *japtvd (em. : japtd Cod.) cdstasatam 
punah || 199 avadhutatanur bhutva prayogam idam drabhet | pasubijasamdyuktam 
V-kdrenaiva bheditam || 200 karsaye tu samddhistho raktaugham raktayd saha 
| tena raktena mantrajnah paripurnakapdlake || 3.201 sugandhakusumair yukte 
tendrgham tu praddpayet \ devindm devadevdya sarvasiddhyarthakdranam || 3.202 
datte 'rghe tu prasiddhyeta trailokyam ndtra samsayah | athavd caiva ii-kdram 
pasubijasamanvitam || 3.203 codayitvd uddnena avadhutatanuh *sadd (corr. : 
saddh Cod.) | nirdcdrena bhdvena pasudeham viset tatah || 3.204 tatrastho 
grahanam kurydt bhutdndm mantracintakah | apdnena tatah sighram svadeham 
pravised budhah || 3.205 pancabhutdni cdkrsya pujayita kapdladhrk | raktena 
prathamd<m> devi<m> dvitiyd<m> mdmsabhaksane || 3.206 trtlyd tvak-ca-bhaksd 
tu caturthi medabhaksand | snehena tarpayed devam pancavyomdntasamsthitam 
|| 3.207 etat te paramam guhyam yogesindm tu pujanam | siddhyartham caiva 
mantrlndm khecaratvajigisundm 'After entering before the celestial Mandala he 
should repeat the Mantra eight hundred times. When [in this way] he has become 
one whose body has transcended all duality he should commence the following pro- 
cedure. In deep meditation he should draw out a stream of the [victim's] blood with 
the [Mantra of] Rakta conjoined with the Victim-seed with u as the [final] vowel. 
The Mantra adept should place fragrant flowers in a skull, fill it with that blood, 
and present it as the guest-offering to the goddesses and Bhairava as the means 
of accomplishing all Siddhis. Alternatively he should propel the letter U combined 
with the Victim-seed up [along the central channel] with the ascending vital energy 
and in the state that transcends convention he should enter the victim's body. Once 
within it the adept should take hold of the gross elements [of the victim's body] 
while meditating on the Mantra and then swiftly return into his own body by draw- 
ing in his breath. When he has drawn them into himself the Kapalika (kapdladhrk) 
should worship [his deities with them]. He worships the first goddess by offering 
her the blood, and the second by offering her the flesh to eat. The third eats the 
skin and the fourth the fat. With the fluid of the body he should gratify the god 
[Kapalisabhairava] who resides beyond the five voids [along the central channel]. 
This worship is the highest secret of the Yogesvaris. [I have taught it] to you so that 
Mantra adepts that seek to master the state of the Khecara may succeed'. See also 
Tantrasadbhdva, ff. 181v5-182r2 (27.1-10); Jayadrathaydmala, Satka 3, f. 184r6 
(Yoginisamcdra 5.40): yasmdtra karmano siddhi raktdkarsanapurvikd \ tarpanam 
devatdndm ca 'For in this [system] the success of the ritual and the gratification of 
the deities requires the extraction of [the victim's] blood'; Tantrdloka 16.35c-51b, 
describing the yogic process in detail; and Netratantra 20, which describes how 
Yoginis extract life-essences from their victims in this way in order to offer them up 
to Mahabhairava and thereby liberate them. 
445 Catuspitha ff. 68v-70r (Guhyapltha, Patala 3) and Bhavabhatta thereon 
(Catusplthanibandha ff. 50v4-52v7); Vajraddka ff. 50r7-52r3 (Patala 21); 
Samputodbhava ff. 78r5-80r6 (Kalpa 8, Prakarana 3); Samvarodaya 5.67-69 and 
19.35c-47. In Tibetan tradition this practice is one of the nd ro chos drug or Six 
Teachings of Naropa (956-1040), commonly known in English as his Six Yogas. 
These have been the object of extensive Tibetan exegesis. For English transla- 
tions of some of these works, including the Chos drug gi man ngag attributed to 


The Saiva Age 

means of assisting the dying and the dead — we have seen a ritualized realiza- 
tion of this in the Mantranaya's funeral ceremony taught by Padmasrimitra and 
Sunyasamadhi 446 — ; and the practice of transferring one's consciousness out of 
one's body to pass into and animate a corpse (parakdyapravesah). 447 

Nor is the adoption of the Vidyapitha's practices restricted to externals. It 
also extended into the domain of Yoga. For one of the most striking features that 
distinguish the Yoginltantras from the Yogatantras and indeed from all that pre- 
ceded them in the history of Buddhism is that they based their inner practice 
on the theory that the body is pervaded and sustained by a network of energy 
channels (nddi), variously numbered, with three pre-eminent: two vertical lat- 
eral channels, lalana and rasand, and a hidden third extending between up the 
centre of the body to the head, called avadhutl or canddll, with Cakras located 
along its course, which was to be awakened and perceived as the means of access 
to the bliss (sahajdnandah, mahdsukham) of enlightened awareness. This Yoga 
of meditation on the channels of the vital energy and the Cakras is not found 
in the transitional Sarvabuddhasamdyoga 448 nor indeed in the Laghusamvara, 

Tilopa, the sNyan rgyud rdo rje'i tshig rkang attributed to Naropa, and the Nd ro 
chos drug gi 'khrid rim yid ches gsum Idan of Tsong kha pa (1357-1419) (Gsung 
'bum, vol. ta, pp. 401-532) see Mullin 1996 and 1997. For Tsong kha pa's detailed 
treatment of this practice of ascent from the body see Mullin 1996, pp. 209-215. 
His sources are those Tantras listed here: the Catuspitha (and Bhavabhatta's com- 
mentary), the Vajradaka, the Samputa (= Samputodbhava), and the Samvarodaya. 
Mullin translates the Tibetan rendering of these titles into English. He identifies 
his 'Mystic Kiss Tantra' as the Caturyoginisamputa. It is in fact the Samputa, the 
work that also appears in this translation as the Sambhuta Tantra, reproducing a 
faulty Tibetan transcription of the same title. Tsong kha pa notes that this prac- 
tice of ascent from one's body (utkrdntih) is a unique feature of the highest (bla na 
med) Buddhist Tantra class (Mullin 1996, p. 209). That is so within the Buddhist 
Tantras; but the source of the practice is the Saiva tradition, whose texts have al- 
ways placed a great emphasis on it both in the Atimarga and in the Mantramarga; 
see Pdsupatasutra 5.30—40; Pampdmdhdtmya 11.54—71 (explaining that pas- 
sage); Skandapurdna-Ambikdkhanda, Adhydya 182; Rauravasutrasamgraha, 
Patala 9; Sdrdhatrisatikdlottara 11.13— 19b; Dvisatika-Kdlottara ff. 2v9— 3r6; Tray- 
odasasatika-Kdlottara ff. 30r9-31r7; Kirana, Patala 59; Matahgapdramesvara, 
Carydpdda, Patala 9; Picumata, Patala 100; Mdlinlvijayottara 17.25-33; 
Tantrasadbhava f. 36rll-vl0 (9.294-321); Tantraloka 28.292-302; and, in 
Java/Bali, Jndnasiddhdnta, chapters 3, 5-7, and 20. 

446 See here pp. 126-128. For the Saiva adaptation of this practice as a means of liberat- 
ing the dying see, e.g., Tantraloka 19.1—56 (sadya-utkrdntidiksd utkrdmani diksa). 

447 Vajradaka f. 51rl-3 (21.19-22). In the Saiva literature see Nisvdsatattvasamhitd 
f. 22v4 (Nisvdsamula 7.20), (>) Svacchanda 7.328c-329b; Picumata f. llv (3.228- 
232b); (5.95-101); f. 356r4-v3 (96.19-35); Tantrasadbhava ff. 181v5-182r3 (27.1- 
11); Mdlinivijayottara 21.9-19; and Tantraloka 28.294-300. This practice too is one 
of the 'Six Yogas of Naropa' (nd ro chos drug); see Tsong kha pa, op. cit. translated 
in Mullin 1996, pp. 215-216. 

448 See also Tanaka 1996, p. 272. 


Genesis and Development of Tantrism 

but it is much developed in the latter's ancillary scriptures such as the Vajradaka 
and Samvarodaya, and elsewhere in the Yoginitantras, notably in the Hevajra, 
the Samputodbhava, the Mahamudratilaka, and the Kalacakra. 449 

The elements of this model are 'purifed through equation' (visuddha-) with 
Buddhist soteriological factors, either newly acquired, such as the twenty-four 
sacred sites or long established in the Mahayana, such as the three bodies of 
a Buddha (nirmdnakdyah, sambhogakayah , and dharmakayah), equated with 
the three principal channels, and Means (updyah) and Wisdom {prajhd), whose 
co-functioning (yuganaddhavdhitd) is the way to liberation, equated with the 
lateral pair. 450 But the basic conception is derived from the Yoga of the Saivas in 
general and the Sakta Saivas in particular. 

The Incorporation of Text-passages from the VidyapItha. In the 
light of this evidence of the pervasive similarities between the Yoginitantras and 
the Saivism of the VidyapItha, and considering the fact that these similarities set 
the Yoginitantras apart from all earlier forms of Buddhism, the reader will not 
be surprised to know that there is also evidence that this tradition incorporated 

That the Yoga of the energy channels was one of the principal features that distin- 
guished the Yoginitantras was asserted by the learned of the Mantranaya itself; see 
Sraddhakaravarman cited here on p. 239; also Mkhas grub rje, rGyud spyi, p. 256, 
11. 6-7: phung khams skye mched kyi mam dag gtso bor ston pa's rgyud yin na pha 
rgyud | rtsa'i mam dag gtso bor ston pa ma rgyud 'If a Tantra principally teaches the 
purification of the Skandhas, Dhatus, and Ayatanas it is a Father Tantra. A Mother 
Tantra principally teaches the purification of the energy channels'. In this pas- 
sage the distinction is between the esoteric Yogatantras (Mahayogatantras, Yogot- 
taratantras) headed by the Guhyasamaja and the Yoginitantras or Yoganiruttara- 
tantras exemplified by the Tantras of Samvara and Hevajra, the two divisions of 
what the Tibetans called bla med kyi rgyud 'the unsurpassed Tantra [class]'. Mkhas 
grub rje's tradition rejects this criterion for distinguishing between the two divi- 
sions on the grounds that there are Yoginitantras (Mother Tantras) that also teach 
the purification of the Skandhas and the rest. That is true. We find this, for exam- 
ple, in the Hevajra (1.7.12; 1.9.6-9, 13-14; 2.2.31-36) and the Abhidhanottara (e.g. 
B ff. 20v5-21rl; f. 26r3; f. 36r3-v6; f. 51r3-4; ff. 69v2-70rl). But that is because the 
second-wave Yoginitantras sought to encompass the tradition of the Guhyasamaja 
by incorporating many of its elements. He does not, we may note, support his argu- 
ment by pointing to the presence of the purification of the energy channels in any 
Father Tantra. From the historian's point of view the distinction that he rejects 
remains accurate in spite of his objections, van Schaik (2008, p. 50) has noted the 
absence of material on the manipulation of the internal energies in the Dunhuang 
manuscripts, which represent Tantric Buddhism up to about the middle of the ninth 

For a comprehensive listing of 'purifying equations' for the principal channels and 
four Cakras (the Nirmanacakra at the root of the navel, the Dharmacakra in the 
heart, the Sambhogacakra in the throat, and the Mahasukhacakra in the head) see 
Jnanodayatantra, p. 6, 11. 1-14 (the four Cakras), and p. 6, 1. 20-p. 7, 1. 9 (the three 


The Saiva Age 

and adapted much textual material from the Saiva scriptures in the process of 
producing its own. 

This is particularly evident in the case of the Laghusamvara and its 
satellites. I have reported and tabulated elsewhere correspondences with 
passages in five Saiva scriptures: (1) the Yoginisamcara of the third Satka of 
the Jayadrathayamala, 451 (2) the short redaction of the Siddhayogesvarlmata — 
a much longer redaction, known to Abhinavagupta, has not come down to 
us — , (3) the Tantrasadbhava, (4) the Picumata (IBrahmayamala), and (5) the 
Nisisamcara, all of which are texts of the Vidyapitha. There are also a few 
correspondences with earlier texts of the Buddhist Mantranaya; 452 but unlike 
those the Laghusamvara's parallels with the Vidyapitha are not short passages 
of one or two verses but detailed and continuous expositions that run in two 
cases over several chapters, amounting in all to some 200 verses out of a total of 

451 The Yoginisamcara, though it comes to us as part of the Jayadrathayamala, has 
very probably been incorporated from another source. This is evident from the reg- 
ister of its Sanskrit, from its style, and from its content. This source may be a text 
closely related to the lost Yoginijalasamvara. For it claims at its beginning to be 
about to explain what has already been taught in that Tantra. Jayadrathayamala, 
Satka 3, f. 169r8 (Yoginisamcara l.l-6b): devy uvaca \\pura tu samvare tantre yad 
uktam paramesvara | Han na (em. : tatra Cod.) jhatam maya deva guhyatantrasya 
vistarat || 2 katham sa bhairavo dehas tvayi deva mahabalah \ katham devyo yajanty 
enam kulas tasam kati smrtah || 3 katham kramam mahagudha<m> caram tasam 
katham vibho \ carusiddhih katham tasam etan me bruhi vistaram || 4 evam akarnya 
devesyavadandmburuhacyutam \ vacomrtam mahddevo bhuyo vacanam abravit || 
5 sadhu sadhu mahabhage sarvajhanarthabhajane | maharahasyam atulam yo- 
ginicaram uttamam || 6 pravaksyami samasena srnusv' ekagramanasa 'The goddess 
said: Paramesvara, I have not understood the teaching that you gave of old in the 
Samvaratantra, because of the great length of [that] esoteric text. What is the 
nature, O god, of your mighty embodiment as Bhairava? How do the goddesses 
worship it? How many are their families held to be? How is the most secret proce- 
dure of their worship? How, O lord, do they rotate? And how is one to obtain the 
sacramental substances for them? Explain this to me at length. Having heard thus 
the nectar in the form of words that fell from the lotus of the mouth of the goddess 
Mahadeva replied and said: I congratulate you, illustrious and worthy receptacle of 
the teachings of omniscience. I shall concisely teach you the incomparable great se- 
cret, the unsurpassed Rotation of the Yoginis. Listen with attentive mind'. The last 
part of the first chapter of the Yoginisamcara gives an account of the many classes 
of female supernaturals as the constituents of the body mentioned in the list of 
questions and ends with the words: ity evam yoganiyamam yoginijala*samvare 
(corr. : samcare D) | yathotpannam tu kathitam *niyogam (em. : niryogam D) srnu 
sampratam (D f. 172v4-5, 1.72c-f) 'Thus I have explained to you the arising of the 
order of the pantheon of powers as [taught] in the Yoginijalasamvara. Hear now 
its application(s)'. See also D f. 199v6-7 (7.124c-125b): uktani ydni karmdni yogini- 
jalasamvare || ayutam japtva tu sarvani karoty eva hi lilaya After repeating the 
Mantra ten thousand times he easily accomplishes all the rites that I have taught 
in the Yoginijalasamvara'. 

452 See here p. 163. 


Genesis and Development of Tantrism 

about 700 with some prose equivalent in length to about 80 more. They teach 
the characteristics by which the initiate may recognize women as belonging 
to various classes of Yogini, Dakini, and Lama, and vocabularies of special 
words and gestures (chommdh) for communicating with them when encountered 
(Patalas 15-24), the rules (samaydh) that bind initiates as they engage in 
post-initiatory caryd (Patalas 26-29), the system of Pithas and other sacred 
pilgrimage centres for wandering ascetics engaged in this practice (Patala 41), 
and the characteristics of the ideal sacrificial victim known as a saptdvartah or 
saptajanmd (Patala 49). 453 

These parallels demonstrate a high degree of overlap with the Saiva 
Vidyapitha in the parts of the text and its satellites that deal with the religious 
discipline (samaydcdrah) of the adherents of this form of Buddhism. Still 
lacking, however, was evidence of textual dependence in those parts that deal 
with that discipline's ritual core. But that gap can now be closed. For since 
publishing those results I have located further evidence in what survives of the 
Vidyapitha's scriptures that this corpus was also the source of substantial parts 
of the Laghusamvara's instruction in this domain. The areas of prescription 
in which this textual dependence has emerged are (1) the daily worship of the 
'Kulika' prescribed in the first chapter of the Laghusamvara, (2) the ceremony 
of initiation before the Mandala through which a candidate becomes qualified 
and obliged to practice the Tantra's rites and observance, which is taught 
from the end of the first chapter to the beginning of the fourth; and (3) the 
ritual procedures for supernatural effects, mostly hostile sorcery, that form a 
considerable part of the work and take the form of fire-sacrifices (homah), and 
the use of the Mantras and the name of the target (sddhyandma) to empower 
substances in various ways and combinations to bring about these results. These 
new parallels are as follows: 

1. The worship of the Kulika: Laghusamvara 1.4-7b (< Herukdbhyudaya 
15.6-10) < Picumata 84.9c-16. 

2. The initiation ceremony: Laghusamvara 1.15-4.1 < 8.3-28 of the Yo- 

3. The ritual procedures for supernatural effects: 

a Laghusamvara, Patala 34 < Picumata 41.1-3, 49.3c-4c, 41.4-7b, 

41.12abc, and 41.15d. 
b Laghusamvara, Patala 35 < Picumata 26.1-2b, 26.41c-44. 

453 p or m y tabulation of these correspondences see Sanderson 2001, pp. 41-47. See 
also Sanderson 1985, p. 214, note 106; Sanderson 1988, pp. 678-679; and 
Sanderson 1994, esp. pp. 92-96. 


The Saiva Age 

c Laghusamvara, Patala 36 < Picumata 26.45c-48b. 

d Laghusamvara, Patala 37 < Picumata 29. lab, 30.1, 29.35, 29.38-48b, 

29.50 [cf. 20.56-57], 29.61ab. 
e Laghusamvara, Patala 50, up to v. 19 (the point at which the earlier 

redaction of the text ends) < Picumata 5.17-18, 5.23c-28, 5.63, 5.67, 


Comparison of the textual parallels reveals that it is the Cakrasamvara cor- 
pus that has adopted and adapted the Saiva sources rather than the other way 
round. For the Buddhist versions abound in instances in which it can be seen 
that Saiva material has been misunderstood, crudely, artificially, and incom- 
pletely modified, or rendered contextually incongruous. The Saiva versions, on 
the other hand, seem to me to be entirely free of signs of textual dependence on 
Buddhist originals. 

Before proceeding to demonstrate this through the presentation and analy- 
sis of examples I wish first to address an objection that has been raised against 
my conclusion. 454 I do so before my analysis because that objection, if it were 
valid, would block in advance the force of all my evidence, being based not on 
contrary analyses of particular parallels but on a perceived characteristic of all 
the materials I have identified. This characteristic is that the Buddhist versions 
are less clear in meaning, less grammatically correct. By concluding that the 
direction of redaction is from Saiva materials to the Buddhist in spite of this 
characteristic I am held to have overlooked or violated the textual critic's maxim 
lectio difficilior potior 'The more difficult reading is to be preferred'. This maxim 
means that when one is confronted by two readings, both of which are plausible, 
one should prefer that which is less easily explained as the result of the alter- 
ation, accidental or deliberate, of the other, provided there is a clearly established 
line of transmission between the sources of the divergent readings. Thus, it is 
implied, the less clear and more incorrect Buddhist versions should be judged to 
have preceded the clearer and more correct Saiva versions on the grounds that it 
is conceivable that a Saiva redactor revised a deficient Buddhist version but not 
that a Buddhist spoiled a superior Saiva version. 455 

What exactly the concept of lack of clarity is thought to cover in this argu- 

454 Davidson 2002, p. 386, n. 105; and Gray 2005, p. 8, n. 19. 

455 In fact it is not clear whether these authors think that the application of this princi- 
ple means that the Buddhist versions cannot be secondary or only that it less likely 
that they are. The second alternative alone would accord with a more fundamental 
principle of textual criticism, namely that there are no hard-and-fast rules because 
every textual problem must be regarded as possibly unique (HOUSMAN 1921, pp. 


Genesis and Development of Tantrism 

ment is unclear; but I assume that the authors had in mind not merely gram- 
matical deviations from the Paninian standard of high scholarship, since those 
are seldom difficult to understand, being characteristic of a particular register 
of the language, but also and principally lack of clarity in meaning caused by 
syntactical incoherence and the like, which is indeed a conspicuous defect in the 
Buddhist versions. Indeed they are sometimes barely intelligible, as is revealed 
by fact that the commentators confronted by these passages offer widely diver- 
gent but equally arbitary interpretations. 456 

Now, the objection that a version which is less clear in this sense must have 
preceded one that is freer of these defects, proceeds from a serious misunder- 
standing of how the rule of the lectio difficilior is to be applied. Firstly, like all 
other 'rules' of textual criticism, it should never be put to work mechanically and 
in advance, without the application of thought to the weighing of probabilities in 
each case; and secondly, it should never be invoked to give precedence to readings 
that are grammatically defective, incoherent, or contextually awkward. 457 Lack 
of clarity is hardly likely to the fault of the original framers of the text-passages, 
who, after all, probably knew what they wanted to say in whatever register of 
Sanskrit they chose to adopt. It is much more likely to be the result of incompe- 
tence and/or carelessness on the part of Buddhist redactors who had difficulty in 
understanding the Saiva texts they were cannibalizing. 

The secondary status of the Buddhist versions is also apparent in another 
deficiency: their greater metrical irregularity. In principle that might be ex- 
plained either as the result of the Saivas' having polished the Buddhist versions 
or as the result of indifference to the preservation of metrical form on the part of 
Buddhist redactors as they adapted metrically correct Saiva materials. But the 
latter explanation is much to be preferred. For, as we shall see, metrical irreg- 
ularity is particularly noticeable in the Buddhist versions at those places where 
the imprint of Buddhism is apparent. 458 

Let us assume, however, that there are indeed readings in the Buddhist ver- 
sions which do not derive from the Saiva parallels that I have identified. Would 
these not refute my conclusion that the Buddhist versions are secondary? No. For 

456 See here p. 216. 

457 This point has been made against Davidson and Gray by Szanto (2008b, p. 218). 
On the principle invoked here, that a 'more difficult reading' must be plausible, see 
West 1973, p. 51: "When we choose the 'more difficult reading' . . . we must be sure 
that it is in itself a plausible reading. The principle should not be used in support 
of dubious syntax, or phrasing that it would not have been natural for the author to 
use. There is an important difference between a more difficult reading and a more 
unlikely reading"; Chadwick 1957, p. 255: "The principle lectio difficilior potior 
does not extend to nonsense, . . . ". 

458 See here p. 207. 


The Saiva Age 

the inference that they would rests on the assumption that I consider that the 
Saiva text-passages redacted into the Buddhist versions were exactly those seen 
in these parallels. In fact I hold that the collation of these parallels with the Bud- 
dhist passages demonstrates that the former are, in most cases at least, closely 
related variants of the passages on which the Buddhist redactors drew, and that 
these passages were accessed in what were probably earlier and less elaborate 
redactions of the works in which I have found the parallels, or else in texts of 
the same corpus which are now out of reach, such as the Yoginijalasamvara, 
the Sarvavirasamayoga, the long version of the Siddhayogesvarlmata, and the 
Pancdmrta. 459 For what survives in the manuscript collections of India and 
Nepal is only a part of what once existed, as we learn both from citations of 
other texts in the works of learned Saiva commentators and from the surviv- 
ing scriptural redactions themselves, which, when listing the canon of texts to 
which they belong, mention many works, such as those mentioned above, which 
have not survived or await discovery 460 My argument, then, is not that these 
Saiva parallels are the direct sources of the Buddhist versions but only that the 
Saiva parallels are close enough to the Buddhist versions to reveal the direction 

459 On these sources see Sanderson 2007, pp. 234-237, footnotes 15-16, and 21-22. 

460 See, for example, the list of Tantras 'venerated by the circle of Yoginis' 
given in the first chapter of the Yoginisamcdra as sources on the matters it 
covers (Jayadrathaydmala, Satka 3, ff. D 170v2-171r3 [1.29-42b]): mula- 
tantram kubjika ca yoginijdlasamvaram | *attasambarandmdnam (ABCE : 
attasasvarandgdnam D) hattadhulis tathdpard || 1.30 caldksaram mahdtantram 
visvakrlddvatdrakam | mahdmdyottaram ndma sarvaviramatam tathd | 
1.31 alamgrdsam mahdtantram *kuhcikodghdtam (em. : kruhcikodghdtam 
ABCDE) eva ca | siddhacakram prakdsam ca patam turam Haihdparam (em : 
yathdparam ABCDE) || 1.32 siddhakaulam mahdjdlam tathd bhairavagahvaram 
| kulagahvarandmdnam kuladdmarabhairavam || 1.33 jhdhkdrakulam atyugram 
tathd siddhdmatam subham \ kdcandmatam evdnyat kusumdlikasamjnitam || 1.34 
siddhayogesvaritantram. trikasdrottaram tathd | picutantram. mahdraudram 
vimalocchusmasamjhitam | 1.35 khadgardvanandmdnam tathdnyam taka- 
mandalam (em. : takamandanam ABCDE) | karoti mundamdldkhyam 
siracchedam bhaydnakam || 1.36 hdhdrdvottaram tantram krodham unmat- 
tabhairavam \ ruruydmalam atyugram tathdnyam rudraydmalam \\ 1.37 
umdydmalam evdnyad gauriydmalam eva ca | skandaydmalam evdnyam 
tathd bhairavaydmalam || 1.38 visnuydmalam eva sydn nandiydmalam eva ca 
| sukraydmalam evdnyac chakraydmalam eva ca || 1.39 kapdlisamatam ndma 
meghanddisvaram tathd | harnsayamalandmanam candogram hdtakesvaram || 
1.40 mahdvdmesvaritantram lahkesimatam uttamam \ lampatddyam ca raktddyam 
tathd hadddmatam param || 1.41 durvdsamatam evdnyam evamddyd hy anekasah 
| ete tantravardh proktd yoginicakravanditdh || 1.42 esu tantravaresv eva tdsdm 
cdram vicdritam. The great majority of these works appear to have been lost. 
Works that have survived with titles listed here are distinguished by bold charac- 
ters. Works here that are known only by citations or as loci of attribution in early 
colophons have been underlined. 


Genesis and Development of Tantrism 

of dependence. It is possible, therefore, that any 'more difficult readings' were in- 
herited from this earlier stratum in the development of the Vidyapitha; and this 
mere possibility is sufficient to invalidate the inference of the priority of the Bud- 
dhist versions. If I am mistaken in my conclusion that the Buddhist versions are 
secondary that will have to be demonstrated by presenting a persuasive contrary 
analysis of the relationship between the Saiva and Buddhist versions based on a 
detailed examination of the particulars I have identified. General arguments of 
this kind, which attempt to settle the matter in advance without engaging with 
the specifics of the parallels, will not suffice. 461 

Having dealt with this objection I can now turn to the evidence. In advance 
of a more thoroughgoing demonstration I consider a few passages here that re- 
veal that the Buddhist redactors were using Saiva materials and enable us to 
see how they did so. 

I have mentioned the entry into the Cakrasamvara corpus of two lists of 
Sakta sacred sites. That found in the Vajradaka, ff. 42rl-43v3 (18.10-60) cor- 
responds very closely in the Vidyapitha to Nisisamcara, ff. 16v-19v (4.6b-5.11), 
both in content and wording. The passage lists twenty-four sacred sites and 
identifies for each its presiding goddess, the high Tantric goddess to whose 
family she is assigned, her weapon (dyudham), the site's sacred tree, and 
a guardian Bhairava (ksetrapalah). 462 The version in the Vajradaka leaves 

461 The same applies to a line of defence that objects to my conclusion in a manner that 
renders even a non-specific engagement with the parallels unnecessary. Confronted 
with the information that such parallels have been claimed some are inclined to 
respond with the question "Why would Buddhists have drawn on Saiva sources?" 
The question is purely rhetorical and somewhat plaintive, implying that since the 
authors of these texts were Buddhists they would surely not have drawn on non- 
Buddhist scriptures. The inference has no force at all, because it invokes a notion 
of the nature of Buddhism and consequently of what Buddhists can or cannot have 
done that is derived from texts other than those of this corpus. No amount of evi- 
dence that other Buddhist scriptures were free of dependence on non-Buddhist texts 
can counter evidence that these Buddhist scriptures were not. 

462 Closely related to the Nisisamcara text is a version seen in Kubjikamata 22.23- 
46, which lacks one of its elements, namely the specification of the high Tantric 
goddesses to whose families these local goddesses belong. Another, somewhat di- 
vergent and giving the sites alone and the points on the body that should be 
empowered by them through nyasah, appears in the Vidyapitha's Madhavakula 
(Jayadrathayamala, Satka 4, f. 124rl-5 [Kalikakule pujanirnayah, vv. 16-22 (fol- 
lowed in Tantraloka 29.59-63 (TA): parts of a Kashmirian redaction of the text 
are cited in Tantralokaviveka on these verses (TAV)]; the procedure of the nyasah 
is put in Paddhati form in Kalikulakramdrcana, f. 22r5-v5 [KKK1): attahasam 
sikhasthane caritram ca karandhrake | *kulagiryam (corr. : kullagirye Cod.) priye 
*karne (corr. : karnnam Cod.) *jayantya (corr. : jayamtya Cod.) *uttare punah 
(conj. [cf. jayantlpithapada vamakarne KKK] : uttaroyane Cod.) | 17 *ujjayanya 
(corr. : ujjayanyam Cod.) tu bhriimadhye prayagam vaktramadhyagam | varanasl 
tu hrdaye sripitham skandhayor dvayoh | 18 kanthadese tu virajam *hy erundya 


The Saiva Age 

this Saiva pantheon and its ancillaries intact, the only major deviation being 
that it has four sites that differ from those in the Nisisamcara. Particularly 
striking in the Vajradaka's version is not only the fact that it transmits all the 
details of this distinctively Saiva religious map, which includes such well-known 
deities as Mahalaksmi of Kollagiri (Kolhapur), Hetuka[bhairava] of Devikotta, 
and Vettada/Vetala of Nagara (Pataliputra/Kusumapura), 463 but also that it 
preserves the classification of the goddesses of these sites as belonging to one 
or other of the families of Rakta, Karali, Candaksi, Mahocchusma, Karala, 
Dantura, Bhimavakta, and Mahabala, information that is revelant only in the 
Saiva context, since these are the four Guhyakas and their attendants that form 
the inner retinue of Kapalisabhairava and Canda Kapalini in the Picumata of 
the Vidyapitha 464 and are not encountered to my knowledge in any Buddhist 

(em. [cf. erundiplthapada | udare KKK] : herumdya Cod.) udare priye | *alampuram 
(Cod. KKK : alipuram TAV : hold TA) ndbhimadhye *samdohailapuram priye 
(Cod. [cf. elapurapithapada medasi KKK] : kandordhve paramesvari TAV) | 19 
kandddhdre tu gokarnam *marudesam (corr. : maruddesam Cod. : marukosam TA) 
bhagdntare | atha medhropari bhadre jhatavyam sddhakena tu | 20 daksine *sak- 
thni (TAV : sakti Cod.) *nagaram (corr. : nagare Cod.) *vdme sydt (TAV : vamesyah 
Cod.) *paundravardhanam (corr. TAV : paudravarddhane Cod.) | vdmaskandhe 
purastlram *prsthapuram (Cod. [cf. prsthapurapithapada daksaskandhe KKK] : 
elapuram TAV) tu daksine | 21 *kudyakesl (TAV : udyakesT Cod.) * jdnumadhye 
(Cod. [cf. kundakesipithapdda jdnumadhye KKK] : daksajdnau TAV) *soparam 
(Cod. : sopanam TA TAV) *cottare (em. [=TAV] : cantare Cod.) smrtam \ *ksirika 
(corr. : kslrikam Cod.) *vamahaste (Cod. [cf. ksTrikapithapada vamahaste KKK] 
tu *mayapurya (corr. : mayapuryan Cod.) tu daksine | 22 amratakesvaram gulphe 
vame rajagrham subham | padadhare tu brahmani kalagnyavadhidharaki. 

463 rp^g name f yjg goddegg of this city is Vettavasim in the Nisisamcara (f. 17v 
[4.43]; em. : vettavasim Cod.) Vetrakacchanivasa in the Kubjikamata (22.37c; em. 
[MSS E and K] : cetrakacchanivasa BCDJG : caitrakacchanivasa Ed.), and Vetra 
in the Kdlikakulakramdrcana (em. : vatrd Cod.). In the Buddhist version we see 
Vettada in the Vajradaka (em. : vettaheti Cod.) and Vetada in the Dakarnava. The 
Vdsavadatta of Subandhu (p. 16, 1. 2 to p. 17, 1. 4) independently identifies her as 
'the Katyayayam called Vetala': kusumapuram . . . yatra . . . katyayani vetalabhidhd. 
We therefore have two phonetically related but semantically unrelated names, one 
meaning the goddess 'who dwells in the thicket of reeds (vetra-)' and the other 'the 
female Vetala', vettada- and vetada- being well-attested variant forms of vetala-. 
I propose that the latter evolved from the former through a vernacular synonym 
*Vettala corresponding to Sanskrit Vetralaya. Cf. Panjabi and Hindi aid from Skt. 
alayah; Panjabi sivala, Maithili and Hindi siwald from Skt. sivalayah; and Panjabi 
dewala from Skt. devalayah. The Mahayanist Mahdsamnipatasutra's Candragarb- 
hasutra, preserved only in a Chinese translation made by Narendrayasas in 566, 
gives in its 18th chapter (Mahasamnipatasutra, chapter 55) a listing of the pre- 
siding deities of 55 places extending from India through Central Asia to China 
(55a-58a [prose]; 59a-60a [verse resume]). The name of the guardian goddess of 
Pataliputra is said there to be Bi-lu-chi or Bi-lu-tuo (Levi 1905b, p. 265). It is 
tempting to see this as a deformation of the same name caused by an inadvertent 
inversion of the last two syllables. But I am not qualified to judge the matter. 

464 See, e.g., Picumata f. 19r2— 3 (4.254c— 256): guhyakddyam tato vaksye namato 


Genesis and Development of Tantrism 

context outside this text-passage and its derivatives. Thus, for example, the 
Nisisamcara (4.10-13), covering Kolagiri (Kolhapur) and Jayanti, reads: 

10 kolagirya<m> mahalaksml karalayonisambhava \ 
kdlarupd sthitd devi dandahastd subhisand || 

11 tasmin ksetre sthitd devi parvatagrasamasrita | 
agniketi ca vikhydtah ksetrapdlo mahdtape || 

12 jayantya<m> danturayoni<r> jvalamukheti visrutd \ 
khadgahastd sthitd devi sarvasattvabhayamkari || 

13 tasmin ksetre sthitd devi nimbavrksasamdsritd | 
mahapreteti vikhydtas tasmin ksetre mahdbalah \\ 

ff. 16v4-17r3 

13a tasmin ksetre corr. : tasmim ksetrd Cod. 

and the corresponding passage in the Vajradaka (18.12-14) reads: 

12 kollagiryam mahalaksml karalayonisambhava \ 

kardlarupd sthitd devi vikrtd cdtibhisand || 

13 tasmin nagare sthitd cogrd parvatagrasamasrita | 465 

varnatas tathd || 255 raktd kardlT *canddkhyd (corr. : canddkhydm Cod.) 
mahocchusmd tathaiva ca | ucchusmatantre ndmdni guhyakdndm na samsayah 
|| 256 kardld danturd caiva bhimavaktrd mahdbald \ guhyakdnucard hy etdh 
kimkaryo 'nukramena tu 'Next I shall explain the [retinue] that begins with the 
Guhyakas, giving their names and colours. In [this scripture,] the Ucchusmatantra, 
the names of the Guhyakas are, without doubt, Rakta, Karali, Candakhya 
(/Candaksi), and Mahocchusma. Karala, Dantura, Bhimavaktra, and Mahabala: 
these are respectively their attendant servants'. The Ucchusmatantra is the 
Picumata itself (f. 185r4: ity ucchusmatantre picumate nddisamcdrapatalah sat- 
trimsatimah). The four secondary goddesses that attend the Guhyakas are also 
called their Dutis. I have not emended canddkhydm, because although Candaksi is 
the standard form of the name there are several other places in this text in which 
the goddess is called Candakhya. 
465 Both the Nisisamcara and the Vajradaka read parvatagrasamasrita (rDo rje mkha' 
gro f. 49r7: ri yi rise mor brten te gnas) 'on a hilltop' here. This is surpris- 
ing because what we expect is a reference to the site's sacred tree, as in the 
parallel expression nimbavrksasamdsritd 'by a Nimba tree' in the next verse. 
It is tempting to emend, therefore to parpatdgrasamdsritd 'in front of a Box 
[tree]', since this is so close to the transmitted reading. However, two consid- 
erations oppose this: (1) in a passage on Kollagiri in the Picumata (f. 7r3-4 
[3.84-87]), which agrees in giving Mahalaksml as the goddess, Agnika as the 
Ksetrapala, and dandah as the weapon, the sacred tree of the site is said to be a 
Vaibhitaka (84 daksinena likhen mantri mahdghoram bhaydvaham \ mahdraudram 
smasdnam tu ndmnd kollagiri tathd || 85 tatra dandam samdlikhya madhye 
vaibhltakadrumam \ ndndvrksasamdkirnam kolldgiryoparis tathd \\ 86 citibhih 
prajvalantibhih samantdt parivdritam \ diksus caiva vidiksus ca bahis tasya 
mahdyase \\ 87 tasyddhastdl likhet padmam astapatram sakarnikam \ agnikam 
ksetrapdlam tu mahalaksmibhaydvaham); and (2) in the Kubjikdmata's parallel 
version of this material Mahalaksml is described as 'residing on a hill' (22.25: ag- 


The Saiva Age 

agnimukheti vikhyatah ksetrapalo vardnanah \ 
lAjvalamukhiti vikhyata \ 

khadgahasta sthitd ghord nimbavrksasamdsritd | 
ksetrapalo mahdkdyo mahavrateti visrutah \\ 

f. 42r2-4 

13c vikhyatah corr. : vikhyata Cod. 14b khadgahasta sthitd em. : khadga- 
hastasthitd Cod. 

Moreover, this Buddhist parallel provides additional evidence of the direc- 
tion of redaction through the state of verse 14. For it lacks the first quarter, 
which contained information vital to the coherence of the passage, namely the 
name of the site over which the goddess Jvalamukhi presides and the goddess of 
the Picumata to whose family she is assigned. As a result of this error, commit- 
ted either by a Buddhist redactor or inherited from a defective Saiva manuscript, 
what was originally the second quarter has become the first. Aware that the met- 
rical cadences required at the end of first and second quarters of a verse in this 
metre are different the redactor has removed the resulting metrical blemish by 
substituting the synonym vikhyata for visruta. But this was not enough, since 
to mend the unmetrical mess that resulted from the omission he would have had 
also to recast the quarters that follow. This was evidently beyond his competence 
or required more effort than he thought necessary. The result is a verse with 
five quarters (a, a, b, a, b) or one and a half verses of which the first half verse 
consists of a prior quarter without the posterior quarter required to complete it. 

As for the four sites found in the Vajradaka's version but not in the 
Nisisamcara, namely Uddiyana, Jalandhara, Tibet, and Malava, there can be 
little doubt that the presence of the third is the work of a Buddhist redactor, 
since Tibet had no religious significance for the Saivas but much for the Bud- 
dhists from the eight century onwards. As for the other three, their presence 
might be explained by assuming that the direct source of the Vajradaka's 
passage was not the Nisisamcara as we find it in its single surviving Nepalese 
manuscript but rather a closely related redaction either within another version 
of the Nisisamcara, such as we find in the paraphrases and citations of a work of 
this name in the Tantraloka of Abhinavagupta and Jayaratha's commentary, 466 

nikena samopetdm dandahastdm nagaukasam \ koldgirye mahalaksmim naumi 
laksmivivardhanim). The hypermetrical reading karalarupa in 12c, which was also 
that of the Tibetan translation (rDo rje rakha' gro f. 49r6: gtsigs pa'i gzugs can), is 
no doubt an error for kdlarupd, echoing kardld in the preceding quarter. 
466 See the paraphrase of the Nisisamcara 's treatment of these twenty-four Sakta 
sacred sites in Tantraloka 15.88-97b and the direct citations in Jayaratha's 
commentary on these verses. These show a list that differs somewhat from 
that found in the Nepalese manuscripts. The latter has Attahasa, Caritra, 


Genesis and Development of Tantrism 

or within some other Saiva source. However, this is improbable in the light 
of the Vajradaka's treatments of all four of these sites. For what they have in 
common is that they deviate from the pattern of the rest of the passage in that 
their presiding goddesses, Mahadevi of Uddiyana, Candalini of Jalandhara, 
Sahaja of Tibet, and Seka of Malava, are not assigned to one or other of the 
eight goddesses of the Picumata. Instead, in the case of the first three the 
redactor has filled in the text at these points by assigning them to the families 
of Guhya (guhyakhyayonisambhava), Soma (somasambhava), and Svayambhu 
(svayambhuyonisambhava), and in the case of the fourth omitting to assign her 
to any deity 467 Why he chose these names is unknown to me. Only one is a 
goddess and not one of them is of any significance in Tantric Buddhism, unless 
the Svayambhu intended is that of the famous Svayambhucaitya of Kathmandu. 
It seems likely that he supplied these names at random in order to maintain the 
compositional structure. In any case, since it would have been an easy task to 
insert names from among those of the eight goddesses that structure his Saiva 
source, it is evident that they meant nothing to him. 

The other list of sacred places appears in Laghusamvara 41.6-15. The 
verses first list these places (6-8b) and then state the classes of Yoginis and 
other female supernaturals said to be present in them, though without covering 
them all. 468 The Saiva source, or rather a later redactional variant of it, is seen 
in the following passage in the Tantrasadbhava: 

Kolagiri, JayantI, Ujjayim, Prayaga, Varana, and Kotivarsa (/Devikotta) (the 
eight Ksetras); Viraja, Erudi, Hatapura, Elapura, Gokarna, Marukesvara, Na- 
gara (Pataliputra), and Pundravardhana (the eight Samdohas); and Parastira, 
Prsthapura, Kundl, Chosmara, Ksirika, Mayapuri, Amratikesvara, and Rajagrha 
(the eight Upaksetras). The list in the redaction known to Abhinavagupta and 
Jayaratha has Prayaga, Varana, Attahasa, Jayanti, Varanasi, Kalinga, Kuluta, 
and Lahula (the eight Ksetras); Viraja, Erudi, Hala, Elapura, Ksirapuri, Na- 
gara, Mayapuri, and Marudesa (the eight Samdohas); and Jalandhara, Nepala, 
Kasmira, Gargika, Hara, Mlecchadigdvaravrtti, Kuruksetra, and Khetaka (the 
eight Upasamdohas). It is striking that this introduces a number of Himalayan re- 
gions, namely Kuluta (Kulu), Lahula (Lahul), Nepala, Kasmira, and also Gargika, 
if that refers to Garhwal. Mlecchadigdvaravrtti 'the pass (?) to the region of the 
barbarians' is also likely to refer to a location in the Himalaya or Hindu Kush. 

467 Vajraddka f. 43rl-2 (18.43): *odydyane *mahadevT (corr. : mahadevi Cod.) 
guhyakhyayonisambhava | vajrasrhkhaladhard devya sughora divyariipini; f. 
43r2— 3 (18.45): jalandhare tu candalini jneya mudra kattarikodyata | soma- 
sambhava mahadevi sarvaisvarya*pradayika (em. : dayika Cod.); f. 43r7-vl 
(18.55): bhotavisaye sahajakhya makaradhvajadharinl | svayambhuyonisambhava 
saumyasya divyariipini; f. 43vl-2 (18.57): malave tu tatha seka mudramud- 
gara*dharini (corr. : dharani Cod.) | sadhakdnam *priyd (corr. : praya Cod.) nityam 
\jasasvini prasasyah\syuh. 

468 A related system of thirty-two sacred sites is taught in Hevajra 1.6.10-19, and, with 
some differences, in Mahamudratilaka, Patala 10 (ff. 17vl-20v5). 


The Saiva Age 

kulutayam aranyese sindhudese nagesvare || 

62 samudrakuksyam saurastre pretapuryam himalaye | 
kancyam lampakavisaye kalinge kausale sthale || 

63 trisakunis tathd caudre kamarupe ca mdlave | 
devlkotte sudharame goddvaryds tate 'rbude || 

64 esu desesu yah kanyah striyo vd klinnayonayah \ 
sarvds tah kamarupinyo manoveganuvrttayah || 

65 sesesu yds samutpannah sdkinyo ghoramatarah | 
sad yoginyah kulutayam aranyese ca matarah || 

66 sindhudese bhaginyas tu nagese kulanayikah | 
samudrakuksyam kampilyah saurastre grhadevatah || 

67 pretapuryam mahakalyo rupinyo himavadgirau | 
kancyam ambah samakhyata lampakavisaye 'mrtah || 

68 kalinge vratadharinyah kausale pisitasanah \ 
cakravakyah sthale proktas trisakunyamarah smrtah || 

69 desadvaye ca sdkinyo nayika viranayika<h> | 

126 yds canyas ca vinirdista raudra bhairavamatarah 
mahamanthanarudras tu tasam mandalandyakah i69 || 

ff. 109v5-110rl, lllvl (16.61c-69a, 16.126) 

62a samudrakuksyam corr. : samudrakuksya Cod. 62c kancyam em. : kamcya 
Cod. 63a caudre corr. : codre Cod. 64a esu em. : esa Cod. 68b trisakunyamarah 
conj. [Aisa Sandhi for trisakunyam amarah] : trisamyamarah Cod. 

The corresponding passage of the Laghusamvara is not present in the in- 
complete Sanskrit manuscript accessible to me, since the folios that contained it, 
covering 38.13c to the end of Patala 44, are among those it lacks. But it can be 
restored with some confidence, except in the matter of the presence or absence 
of a few particles, by combining the evidence of the Tibetan translation, 470 the 

469 rp^g f ac ^ ^ a ^ ^g ^ ex £ f 09a]-, an( j \26 are contiguous in the Buddhist version indi- 
cates that the Saiva text on which it drew was not the Tantrasadbhdva, at least not 
in its surviving redaction, but an earlier source to which 69c-125, which contain 
a further, much longer list of Sthanayogims and their classification as belonging 
to the families of one or other of the seven Mothers (sapta matrkulani), have been 
added. The alternative, that the Buddhist redactor removed this section because he 
had no use for this list and its scheme of classification, is not impossible. However, 
it seems unlikely that in that case he would have taken the special trouble of re- 
taining 126. It is not needed to complete the sense and proved awkward to integrate 
because he had it in what was evidently an already corrupted form. 

470 bDe mchog nyung ngu, f. 238vl— 5 (= Laghusamvara 41.6—15): kuluta dang dgon pa 
dang | si ndhu'i yul dang grong khyer dbang \ gser gyi gling dang sau ra sta | de 
bzhin lha yi khyim dang ni \ yi dags grong dang kha ba'i gnas \ ka nci 'am la mpa ka 
yi yul | ka li ngga dang ko sa la \ tri sha ku ne o tre dang \ ka ma ru pa ma la wa lha 
mo'i mkhar dang ra ma'i dbang \ go da ba ri a rbu da \ au dya na dza la ndhar dang 


Genesis and Development of Tantrism 

lemmata in the surviving Sanskrit commentaries, and a rewriting of parts of the 
passage in the Vajradaka: 411 

41.6 kulatayam aranye ca sindhudese nagaresvare | 
suvarnadvipe saurastre tatha ca grhadevata 
pretapuryam himdlaye || 

7 kdncydm lampdkavisaye kalihge cfaiva] kosale | 
trisakunis tatha odre kamarupe [ca] malave || 

8 devikotte ramesvare godavarydm [tathd]rbude | 
oddiyanajalandharapulliramalayadisu \\ 

9 etesu desesu kanya yd vlradvayavyapini | 
sarvas tah kamarupinyo manoveganivrttayah || 

10 sad yoginyah kulatayam marudese ca matarah | 
sindhudese [ca] lamas tu nagare kulanayikah | 

11 lampake saurastre kuladevatah \ 
pretapuryam mahdkalyo dakinl saha rupinl || 

12 himagirau kdncydm sabalikah \ 
pancalavisaye grhadevata || 

13 kalihge vratadharinyah kosale pisitasanah | 
pretapuryam vajradakyah sthalesvare || 

14 trisakunyam [ca] amarah pulliramalaye 
kanakagirau antyajah striyah sahasrany ekavimsatih || 

\pu Hi ra ma la ya sogs | yul 'di dag gibu mo gang | dpa' bo gnyis med rnal 'byor ma 
| de kun 'dodpa'i gzugs can te \ yid kyi shugs kyis 'jug pa yis | rnal 'byor ma drug ku 
lu tar | myang ma yul na ma mo mams \ si ndhu'i yul na la ma ste \ rigs kyi gtso mo 
na ga rar | la mpa ka dang sau ra stra | rigs kyi lha mo mams yin no | yi dags grong 
dang nags chen par \ mkha' 'gro ru pi ka ru bcas | kha ba'i ri dang ka heir ni | byis 
bcas ma ru bshad pa ste | pa hca la yi yul dag na | khyim gyi lha mo ka li nggar | 
brtul zhugs 'dzin pa mams yin no \ ko sa lar ni sha za ba \ yi dags grong du de bzhin 
du | rdo rje mkha' 'gro sbom dbang phyug \ tri sha ku ner du ma skyes ma \ pu li ra 
ma la ya de bzhin \ gser rir sme sha can rigs skyes \ bud med stong phrag nyi shu 
gcig | lhag ma gzhan dag ji snyed pa | dpal Idan he ru ka yi ni | 'khor lo'i rnal 'byor 
ma yin no | he ru ka dpal sbyor ba che de yi dkyil 'khor gtso mo yin . 
471 Vajradaka f. 41v3-6 (18.3c-10b): sad yoginyas tu sadhakah mlecchabhasam tu 
bhasitam | 18.4 kulatayam tu marudese ca yd matarah || sindhau ca nagare *ca 
yah (corr. : carya Cod.) kulanayikah | 18.5 lampake saurastre ya<h> kuladevatah 
| himagirau *kahcyam yah sabalikah (em. : kahcayam yd balika Cod.) | 18.6 
pancala grhadevatayam yd kanya sahajarupinl \ kalihge *kosale (corr. : kausale 
Cod.) caiva vratadharinT *pisitasana (em. : pisitasina Cod.) | 18.7 pretapuryam 
trisakunau ca sthulesvarl khanda*rohika (em. : rohita Cod.) sthita \ *purnagirau 
(corr. : punnagirau Cod.) jalandhare candalajah striyah | 18.8 odre kamarupe 
ca mahakanyah devikote ramesvare ca yd kanya mata | *goddvaryam arbude ca 
(corr. : goddvarydmbude va Cod.) dakinl paramesvarl \ 18.9 suvarnadvipa<m> 
*yathoddistam (corr. : yathodhistam Cod.) udyayanam tathaiva ca \ etesu desesu 
yd kanya vlradvayavyapini | 18.10 sarvas tah kamarupinyo * manoveganivrttayah 
(corr. : manovegonivrttayah Cod.). 


The Saiva Age 

15 any dpi sesas ca yavatyah sriherukasya yogini | 
mahamanthana tasam mandalanayika \\ 

The words within square brackets are purely conjectural 

Testimonia: BhBh = Bhavabhatta ad loc; DG = Devagupta ad loc; JBh = Jaya- 
bhadra ad loc; KP = Kambalapada ad loc; Tib. = bDe mchog nyung ngu; VD = 
Vajraddka f. 41v3-6 (18.3c-10b). 

LEMMATA: 6a kulatayam ityddind BhBh • aranyam marubhumih JBh 6d 
grhadevateti saptamilopdt BhBh 8a drano rdmesvarah JBh 8cd oddiydnajdlan- 
dharapulliramalayd ddibhutd yesdm ta oddiydnajdlandharapulliramalayddayo 
'rbudddayah BhBh; pulliramalayo na nirdistah JBh 9ab etesu desesu KP, BhBh, 
VD • yd kanyd virddvayavydpini VD, BhBh, KP; bu mo gang dpa' bo gnyis med 
rnal 'byor ma (yd kanyd virddvayayogini) Tib.; 9c kdmarupinya iti BhBh, VD 
9d manoveganivrttaya iti BhBh, KP, VD 10a sad yoginyah BhBh, KP, JBh, VD 
10b marudese BhBh, KP • mdtdrd iti BhBh; mdtarah kdkdsyddydh JBh lOab 
kulatayam marudese ca mdtaretyddi KP, VD 10c lamas tv iti JBh; lama iti BhBh 
lOd kulandyikdh JBh, BhBh llab la mpd ka dang sau rd stra Tib.; lampdke 
saurdstre yd<h> kuladevatdh VD; lampdydm saurdstre kuladevatdh BhBh; lie 
mahdkdlo mahdbhairavah lied pretapurydm mahdkanyd ddkinlsaharupiniti 
BhBh; ddkinibhir iti sahdrthe trtiyd \ kimbhutdbhih saha \ rupinyah \ rupinity 
anyd rupinyas cumbikdsabdlikdprabhrtayah prthagbhutdh saha rupinibhir iti 
drastavydh 12ab himagirau kdncydm sabdlikd iti BhBh 12cd pancdlavisaye \ 
grhadevatd grhadevatdydm BhBh; pahcdla iti JBh 13a ka li nggar | brtul zhugs 
'dzin pa mams yin no (kalinge vratadhdrinyah) Tib.; kalihge ca vratadhdrinyah 
BhBh 13b kosale pisitdsandh BhBh 13cd pretapurydm vajraddkinyah BhBh 
14bcd pulliramalaye kanakagirdv iti \ ihdntyajdh striyah \ sahasrdny ekavimsatir 
iti bdhulyasucandrtham BhBh; sahasrdny ekavimsatir iti KP 15ab sesdnyesu 
yavatyah sriherukacakrayoginltyddi KP, BhK (lhag ma gzhan dag ji snyed 
pa | dpal Idan he ru ka yi ni | 'khor lo'i rnal 'byor ma yin no), DG (lhag ma 
gzhan mams ji snyed pa | dpal Idan he ru ka yi ni | 'khor lo'i rnal 'byor ma yin 
no); sesdnyesu hi ydvantya iti | sriherukasya yoginiti prathamdbahuvacanalope 
BhBh; anyd api sesas ca devatyah sriherukayoginyah JBh (cf. DG: lha mo gzhan 
dag ji snyed pa | dpal Idan he ru ka yi ni | zhes bya ba la sogs pa smos so | ji 
Itar zhen | he ru ka yi sbyor chen las | de yi dkyil 'khor gtso mo yin | zhes bya 
ba la sogs pa la) 15cd mahamanthana iti sriherukasya manthdnayogydh 
tasam iti nirdhdrane sasthi | mandalanayika iti tricakravartinyas caturvimsatir 
ddkinyah JBh; mahdmanthdnam prajhopdyasvarupatvam updyo vd | tendnvitah 
sriherukah prajndriLpah tasya sambandhinindm tasam madhye mandalanayika 
vajravdrdhi samdpanneti bhdvah \ mahdmanthdnam nirmdnam nirvibhaktikam 
| tasam nirmdnam sriherukenaiva sampddyam yatah | sriherukamahdmudrd- 
mandalandyiketi kecit BhBh 

In the Buddhist version the total of twenty-one sites has been raised by the 
addition of Oddiyana, Jalandhara, and Pulliramalaya at the end of the first sec- 
tion (8cd). The reason for the addition is not made explicit in the Laghusamvara 
itself; but the fourth Patala had listed twenty-four Yoginis from Mahavlrya to 
Pracanda; 472 and in the ritual system followed by the commentators and the 

472 Laghusamvara f. 4v4— 6: *tato (JAYABHADRA : tatah Cod.) dakinyo bhuvanani 
vijrmbhayanti | 4.1 mahdviryd cakravartini mahdbald suvird cakravarmini \ 


Genesis and Development of Tantrism 

corpus of explanatory Tantras the sacred sites, as we have seen, are likewise 
twenty-four because each is the location of one of these Yoginis. We have evi- 
dence of two stages in the modification of the text that produced this result. For 
the earlier redaction, attested by Jayabhadra, states that Pulliramalaya is not 
mentioned in this passage but must be understood to be included. 473 It is clear 
then that his text mentioned only Oddiyana and Jalandhara in addition to the 
twenty-one of the Saiva source. Jayabhadra does not cite the actual wording 
of the insertion, and no other indications allow us to establish it. However, it 
is unlikely that the redactor took the trouble of stretching his interpolation of 

saundini khandarohd cakravega khaganana \\ 2 haya*karnd (corr. : varnnd Cod.) 
subhadra ca *sydmddevl (corr. : sydmdthavT Cod.) tathaiva ca | surdbhaksi vdyuvegd 
tathd mahdbhairavd || 3 airdvati drumacchay a lahkesvarl kharvarl tathd | viramati 
mahdndsd prabhdvati caiva canddksT pracandd ca sddhakah || 4 etdh siddhds tu vai 
purvam caturvimsati ddkinyah. This list too has parallels in the Vidyapitha, though 
I have found only partial matches. Thus the Yoginisamcdra of Jayadrathaydmala, 
Satka 3, gives the following list of twenty-four Yoginis whose names when ut- 
tered draw in the Smasanabhutas (f. 202r5-7 [9.58-61]): sarabhdn<an>d suvira 
ca vajribhd *rdsabhd (conj. : rdsibhd Cod.) tathd | *cakravarti (corr. : cakravarti 
Cod.) ca *saundi (em. : paundi Cod.) ca khadgakarnd mahdtapd || 59 cakravega 
mahdydmyd subhadra gajakarnikd | card vai somadevT ca gavdksT vayuvegaga 
|| 60 airdvati mahdndsd damstrdlT ca sukarkasd | vedhani ca tathd bhattd 
drond kdkenakd tathd || 61 yatra ndmdni yogindm uccdryante mahdtape | tatra 
smasanabhutas ca sdmnidhyam yd<n>ti tatksandt. The eight names in bold char- 
acters are those that are among the twenty-four of the Laghusamvara. Compare 
also the names Sarabhanana, Khadgakarna, Gajakarnika, and Somadevi with the 
Laghusamvara 's Khaganana, Hayakarna, and Syamadevi. The names of four of 
the Laghusamvara's Dakinls are found among the fourteen inner goddesses of the 
Picumata, i.e., the four Guhyakas, their four Dutis, and the six Yoginis, namely 
Candaksi (the third Guhyaka), Mahabala (the fourth Duti), and Cakravega and 
Mahanasa (the fifth and sixth Yoginis). For the first eight see 4.254c-256 cited 
here p. 193. For the six Yoginis see f. 19r3 (4.257): krostuki vijayd caiva gajakarnd 
mahdmukhi | cakravega mahdndsd sad yoginyah praklrtitdh. Suvira appears 
in Kubjikdmata 21.45c and Matasdra f. 138rl, Khaganana as one of the eight 
Saktasiddhas of the Kalikula/Krama, Lankesvari in Matasdra f. 81rl as one of 
eight Yoginis in a variant of the inner retinue of the Picumata, and Prabhavati in 
Kubjikdmata 11.115a and 12.23b. 
473 See here p. 158. Kanha, Yogaratnamdld on Hevajra 1.7.12, identifies Pulliramalaya 
with Purnagiri and that appears in its place in listings of these sacred places in 
later texts of the Cakrasamvara cycle, as in Samvarodaya 9.14. In the treatment 
of the thirty-two sacred sites of the Hevajra system in the tenth Patala of the 
Mahdmudrdtilaka we find Purnagiri and Pullira denoting the same place (f. 17r5- 
vl: odiydnam pltham dkhydtam pitham jdlandharam smrtam | pitham purnagiris 
caiva kdmarupam tathaiva ca . . . f. 18rl-2: sirasi sthitam vajrapltham sikhdydm 
jddisamjnitam \ pulliram mastake jneyam bhrumadhye kdmarupakam). On the 
location of Purnagiri, in the Deccan, see Sanderson 2007a, pp. 298-299. In Sakta 
Saiva sources it is one of the principal Pithas and is often referred to, but never 
under the name Pulliramalaya/Pullira. 


The Saiva Age 

the names of these two places to fill a whole line (8cd). 474 The later reading, 
oddiyanajalandharapulliramayadisu, attested by the Tibetan translation and 
by the lemmata in the commentaries of Bhavyakirti and Bhavabhatta, supplies 
the missing Pulliramalaya and, incidentally, is an almost metrical half-verse: its 
first half (oddiyanajalandhara) is unmetrical, but the second is not, and together 
they provide the required total of sixteen syllables. As for the meaning of the in- 
sertion, ordinary usage suggests that it is 'Oddiyana, Jalandhara, Pulliramalaya, 
and others'. But that would not sit well with the closed list of twenty -four Yo- 
ginis to which the sacred places were required to correspond. Thus it has been 
interpreted by Bhavabhatta to mean 'beginning with Oddiyana, Jalandhara, and 
Pulliramalaya', this compound with its locative plural ending being read as qual- 
ifying the twenty -one sites, each listed in the common text with actual or virtual 
locative singular endings. Thus we have twenty-four Yoginis in twenty-four sites. 
All that was needed to make this fit the system known to the commentators was 
to claim that the Laghusamvara is deliberately concealing the true order of the 
items, both the names of the Yoginis in Patala 4 475 and the names of the sacred 
sites in Patala 41. For in their system that order is not Oddiyana, Jalandhara, 
and Pulliramalaya followed by the twenty-one from Kuluta to Arbuda, as the 
Laghusamvara itself indicates, but the added three in reverse order followed by 
the twenty-one in reverse order, with the order of the Yoginis also reversed, so 
that the true sequence is from Pracanda in Pulliramalaya to Mahavirya in Ar- 
buda. 476 

474 The frequent deviations from correct metrical form in this corpus create the im- 
pression that the redactors were largely indifferent to this aspect of composition, 
happily inserting and deleting without feeling the need to rewrite the result to con- 
form to the rules of the Anustubh metre. The alternative, that they lacked not 
the inclination but the ability to do so, seems to me less likely. In the texts of the 
Saiva Vidyapitha, even when the Sanskrit is of a register well below that of the 
learned, the metrical structure is generally sound. Indeed since we find forms from 
both learned and scriptural (Aisa) registers used in the same texts it seems that by 
drawing on both the redactors were not only asserting that their compositions were 
divine rather than human utterances but also making the task of metrical compo- 
sition easier for themselves by using an Aisa form that fitted the metre when the 
Paninian would not, as, for example, in the case of the not infrequent use of Aisa 
genitives plural in -dm in place of the Paninian -dndm. 

475 On the passage listing the twenty-four Yoginls/Dakinls in Patala 4 Jayabhadra com- 
ments (Cakrasamvarapanjikd, p. 115): tricakravyavasthitdndm ddkinindm prthak 
prthah ndmdni kathyante | mahdvlryetyddind vilomena kathitam 'The names of 
each of the Dakinis that occupy the three circuits are now taught. This has been 
done in the reverse order, beginning with Mahavirya [and ending with Pracanda]'. 
The order in which Mahavirya is the last and Pracanda the first, the order of their 
ritual application, is, however, indicated later in the text, in f. 35r7 (48.13): yoginyah 
pracandddayas tathd. 

476 Bhavabhatta, Cakrasamvarapanjikd, f. 126vl-3 (Ed. p. 547): oddiydnajdlandhara- 


Genesis and Development of Tantrism 

Most of the few other differences between the version in Laghusamvara 
41 and that seen in the Saiva source are of little significance. But there is 
one that is more revealing. The Tantrasadbhdva has Sthala between Kosala 
and Trisakuni (16.62c-63b: kancyam lampakavisaye kalinge kausale sthale 
| trisakunis tathd caudre kdmarupe ca mdlave), whereas the Laghusamvara 
lacks it (41.7: kancyam lampakavisaye kalinge cfaiva] kosale \ trisakunis 
tathd odre kdmarupe [ca] mdlave), and instead between Saurastra and Pre- 
tapuri has Grhadevata (41.6: kulatdydm aranye ca sindhudese nagaresvare 
| suvarnadvlpe saurastre tathd ca grhadevata pretapuryam himdlaye), 
which the Tantrasadbhdva lacks (16.61c-62b: kulutdydm aranyese sindhudese 
nagesvare \ samudrakuksydm saurastre pretapuryam himdlaye). Two features 
are immediately obvious here. The first is that the additional words tathd ca 
grhadevata have been added to an otherwise metrically correct verse with the 
result that it has five Padas rather than the required four, with the fourth and 
fifth both with the cadence restricted to the second and fourth Padas of the 
Anustubh, thus crudely violating the required metrical alternation of evenly and 
unevenly numbered Padas that is hallmark of this metre. The second is that 
Grhadevata, meaning 'household deity' is a most implausible place name. The 
key to the mistake, which became a permanent part of the ritual system of the 
Cakrasamvara cycle, is in the second part of the passage in the version of the 
Tantrasadbhdva, which tells the reader the classes of supernaturals that are 
present in the sacred sites. For there grhadevatdh 'household deities' are said to 
be present in Saurastra in a verse in which the items Saurastra, grhadevatdh, 
and Pretapuri are stated in that order (16.66c-67b: samudrakuksydm kdmpilyas 
saurastre grhadevatdh \ pretapuryam mahdkdlyo rupinyo himavadgirau 'In 
Samudrakuksi Kampilis, in Saurastra Grhadevatas, in Pretapuri Mahakalis, 
in Himalaya Rupinis'). Evidently the redactor has read the sequence saurastre 
grhadevatdh pretapuryam as though these were three sites rather than one site 
followed by its resident supernaturals and another site. Probably his manuscript 
read grhadevata rather than grhadevatdh and he took it as a stem-form to be un- 
derstood as locative, a licence of kind seen elsewhere in both the Laghusamvara 
and its Saiva sources, as, apparently, in the unmetrical insertion that this 
error prompted: suvarnadvlpe saurastre tathd ca grhadevata pretapuryam 
himdlaye. Bhavabhatta duly comments on the occurrence of grhadevata in that 

pulllramalaya adibhuta yesam ta oddiyanajalandharapulllramalayadayo 'rbuda- 
dayah kulatantah | bhavas cayam *puiliramalayam adim (Cod. : pulllramalayadim 
Ed.) krtva jalandharauddiyandrbudadisu santity upadesartham vyatikrama- 
nirdesah | etena mandate sarire ca pulliramalayadisu yogininyasah kathitah; 
ff. 126v6-127rl (Ed. p. 547) pulliramalayadisu pracandddaya OM KARA KARA 
PRACANDE HUM HUM PHAD ityddimantrajd bhavyah. 


The Saiva Age 

part of the passage with the words grhadevateti saptamilopat '[We have the 
form] grhadevata [here] because zero has been substituted for the ending of the 

The direction of redaction is also unmistakeable in the passage of the 
Laghusamvara (1.15-4.1) (B) that prescribes the ritual of initiation. This has 
evidently been redacted on the basis a Saiva source of which an expanded variant 
is seen in 8.3-28 of the Yoginisamcara (A) redacted in the Jayadrathayamala: 

8.3 girigahvaraguhyesu 


1.15 girigahvarakunjesu 

mahodadhitatesu ca 
adisiddhe smasane vd 
alikhen mandalam subham 

mahodadhitatesu vd 
adisiddhe smasane ca 
tatra mandalam alikhet 

iti herukabhidhane 
mandalavatarapatalah prathamah 

8.4 smasanabhasmana misram 
kapilagomayam subham 
tena bhumim pralepayet 

8.5 smasanabhasma samgrhya 
smasane 'stadalam subham 
smasanangaracurnam tu 

trirekham mandalam likhet 

8.6 ekahastam dvihastam vd 
caturastakaram tatha 

Cf. B 2.12cd 
sutrayed rudhiraktena 
savasutrena sutradhrk 
Cf. B 2.11cd. 

8.7 akrodhano sucir dakso 
deary o jnanaparagah 

2.1 tatra panagomayena 
mandalabhumi pralepayet 
smasanabhasmana yuktam 

2.2 upalipya tato bhumim 
tatra mandalam drabhet 
smasanam tu samacaret 

2.3 cityahgaracurnena 
alikhen mandalam divyam 
acaryah susalaksanah 

2.4 samyagjhanatantrajhah 
akrodhanah sucir dakso 
yogajho jnanaparagah 

2.5 kapalakrtamurdhajah 


Genesis and Development of Tantrism 

8.8 pancamudravratadharo 
bhairavahgair vibhiisitah 
samantat parivestitam 

8.9 alikhen mandalavaram 

caturasram caturdvaram 

madhye padmavibhusitam 

8.10 astapatram tu tat padmam 
karnikadhisthitam subham 
tasya madhye nyased devi 
bhairavam bhlmavikramam 

8.11 daksindbhimukham diptam 
bhlmarupam bhaydvaham 

tasyagratah sthita devi 

aghora ghoravikrama 

8.12 bhairavabhimukham kruddham 

raudrarupam nyaset tatah 

sambhavan matrair vibhusitagatrah 

mudramantrair alamkrtam 

2.11 alikhen mandalam ghoram 

tato mrtakasutrena 
maharudhirarahjitena va 
Cf. A 8.6cd 

2.12 sutrayen mandalam ghoram 
herukasya param puram 
ekahastam catur astam ca 

Cf. A 8.6ab 

caturasram tu samantatah 

2.13 caturdvarasamakirnam 
vicared dvigunam mantri 
yajed dakinljalasamvaram 

2.14 tasya madhye pratisthapya 
sapatram karnikojjvalam 
puskarais ca kesaranvitam 

2.15 karnikayam nyased viram 
mahdbhairava bhlsanam 
tejaskam tu sudlptahgam 

2.16 kapalamalabharanam 
divyam trinetram caturmukham 
hasticarmavaruddham ca 

2.17 khatvahgakrtahastam tu 
tasyagratah sthitam devim 
vajravarahlm sughoram 

2.18 mahabhairavabhimukham krtva tu 
trimukhlm raudrarupinlm 

8.19c tatah sisyan pravesayet 
sopavasan sucin snatan 
arcayed uttaramukhdn 
8.20 kapalena sirah sprstva 
samputam hrdaye nyaset 
khatvahgena tu sarvahgan 
alabhet putrakasya tu 


The Saiva Age 

8.21 agrato vadayed ghantam 
patahlm damarum tathd 

vastracchannamukham devi 

puspdhjalidharam tathd 
8.22 pradaksinikrtya puram 

daksindmurtim dsritah 

tato ddvdpayet puspdn 

devasyopari putrakam 

8.23 yasmims tat patate puspam 

tat tasya kulam ddiset 

tilakdn tesu kdrayet 

8.24 raktena darsayet tasya 
mukham udghdtya mandalam 
yad yasya devatdsthdnam 

tat sarvam tasya darsayet 

8.25 samaydh srdvayitvd tu 
pranipatya puram guroh 
susrdvya purvavidhind 

8.26 gurum sampujya vidhivad 

3.2 ghantdnddam dlambya 
puspadhupair alamkrtdm 
ghantam vddayet susvardm 
patahikdm vdpi sddhakah 

3.3 hdhdkdram ca kdrayet 
evam vidhivat pujya 
mandalam sarvakdmikam 

3.4 samchddya patavastrena 
mukham tesdm tu putrakam 
puspapurndnjalim praksipet 

3.5 pradaksinam ca tatah krtvd 
sddhakah susamdhitah 
pravesayet tat puravaram ramyam 
daksindmurtim dsritya 

3.6 puspdhjalin tatah ksipet 


yasmin patati tat puspam 

kulam tatra vinirdiset 

3.7 srlherukddipitha darsayet 
tatah puj ay en mudrdm 
dcdryah susamdhitah 

3.8 sisydndn tu dvitiye ahani 
raktena trijaptena 
tilakam tasya kdrayet 
mukham udghdtya sisyam 
darsayen mandalam tatah 

3.9 yad yasya devatdsthdnam 
tatra tarn darsayet samyak 

pranipatya tatah pascdd 

3.11 tatas tu gurave dadyat 

pragrhya kulajdn mantrdn 
vratdms ca samaydms tathd 
8.27 tdvad drddhayed devi 
yoginyo mdtaro gurum 

matrdutyo vratams caiva 
ydvadantam kramena tu 

3.15c tatas tasya tusyanti 

ddkinyo yogamdtardh 
ddkinyo lamayas caiva 
khandarohd tu rupini 


Genesis and Development of Tantrism 

4.1 tato dakinyo bhuvanani 
vijrmbhayanti mahdviryd 

8.28 drddhanakramdd ydti 
bhairavibhuvand devi 

sarvasaktibhir dvrtah 

Apparatus Criticus of A 

Codd. : A ff. 286v2-; B ff. 182r4-; C ff. 166v3-; D ff. 200r2-; E ff. 183v7-. 

8.3c ddisiddhe smasdne B : ddisiddhai smasdnair ACDE 8.7a akrodhano em. 
: sakrodhano Codd. 8.10d bhairavam em. : bhairavT Codd. 8.1 Id vikramd 
em. : vikramdn AC : vikramdm B : vikramdt DE. Cf. Picumata 1.2d: aghori 
bhimavikramd 8.19d uttardmukhdn em. : uttardmukham C : uttardmmukham 
ABDE 8.20b samputdm corr. : samputd Codd. 8.20c sarvdhgdn em. : sarvdhgd 
ACD : sarvdhgo B 8.21a vddayed conj. : vddaye Codd. 8.21b patahim em. : 
pataho Codd. • damarum em. : damaras Codd. 8.21d dharam em. : varam 
Codd. 8.22d putrakam em. : putrakah Codd. 8.25a samaydn em. : samayam 
Codd. • srdvayitvd B : srdvayitvds Codd. 8.25d samsiddhaputrakdnvitam conj. 
: samsiddham putrakdmvitam A : samsiddham putrakdmcitam BCDE 8.28a 
drddhana conj. : aropand Codd. • kramdd ydti conj. : kramaprdpti Codd. 

Apparatus Criticus of B 

Cod.: f. 2r3-. TESTIMONIA : AbhU = Abhidhdnottara 46.10-57 (A f. 146r6- [<La- 
ghusamvara 2.1-]); BhBh = Bhavabhatta ad loc; BhK = Bhavyakirti ad loc; IBh 
= Indrabhuti ad loc; JBh = Jayabhadra ad loc; SV = Sasvatavajra ad loc; Tib. = 
bDe mchog nyung ngu; VV = Viravajra ad loc. 

1.15c ddisiddhe BhBh : ddisiddha Cod. 2.1a tatra pdnagomayena Cod. AbhU, 
BhBh, SV (chu dang ba byung blangs 'water and cow dung') : Hatrdpdtagomayena 
Tib. (der ni lei ba ma Ihung bas), BhK (de la lei ba ma Itung bas) 2.1b prale- 
payet Cod., AbhU : upalepayet BhBh 2.3a cityahgdra BhBh : citydhgdra 
Cod. : cityahgdraka AbhU 2.3b samyutam conj. (= AbhU); cf. Picumata 
5.116cd: kdkavista samdddya smasdnestakasamyutam) : samyuktam Cod. 2.4a 
samyagjndnatantrajnah Cod., BhBh : samyagjndnesu tattvajnah AbhU 2.4c 
akrodhanah JBh AbhU : akrodhas ca Cod. BhBh 2.1 Id mahdrudhiraranjitena 
vd Cod., Tib. (de nas sems med srad bu 'am | ru di ra ni chen pos brlan) 
mahdrudhirdnjitena vd BhBh : mahdrudhiraranjitam AbhU, Tib. 2.13d yajed 
JBh : japed Cod. : pujayed BhBh, Tib. (mkha' gro dra ba'i bde mchog mchod) 
2.17c tasydgratah sthitdm devim JBh, BhBh, Tib. {de mdun gnas pa'i lha 
mo ni) : tasydlihgatdsthitd devi Cod. 2.18a mahdbhairavdbhimukhdm krtvd 
tu JBh : mahdbhairavdbhimukhim AbhU, W (rab 'jigs byed che la phyogs) : 
sriherukdbhimukhdm krtvd tu Cod. BhBh : *mahdsriherukdbhimukhim Tib. 
(he ru ka dpal che la phyogs) : * sriherukajndndbhimukha- (he ru ka dpal ye 
shes phyogs ni IBh 3.2c vddayet Cod. : nddayet BhBh 3.3a pujya BhBh 
sampujya Cod. 3.4b putrakam em. [Aisa gen. pi.; =AbhU] : putrakdn BhBh 
putrakdndm Cod. 3.7a sriherukddipitha BhBh (sriherukddipitheti dvitiydlope) 
sriherukddim pithan Cod. 3.7bc tatah pujayen mudrdm dedryah susamdhitah 
BhBh, Tib. (de nas slob dpon legs par ni | mnyam par bzhag ste phyag rgya 
mchod): tatah pujayen mudrdedryah susamdhitah Cod. : tato hi pujayet mudrdm 
dedryah susamdhitah AbhU : tatah pujayen mudrdm svamudrdm susamdhitah 
JBh 3.9a yad yasya JBh, BhBh : yo yasya Cod., AbhU. 


The Saiva Age 

Here we see several tell-tale signs. In the Buddhist version the disciples 
undergoing the initiation are referred to as putrakah (3.4ab: samcchadya 
patavastrena mukham tesam tu putrakam 'Having covered the faces of those 
disciples with a piece of cloth'), a term that is standard in this technical sense 
in the Saiva literature but to my knowledge appears with it nowhere else in 
Buddhist Tantric sources. 

In 2.15 the installation of the main deity in the centre of the initiation 
Mandala is described as follows: karnikayam nyased vlram mahdbhairava 
bhlsanam 'On the pericarp [at the centre of the lotus diagram] he should install 
the terrifying Vira Mahabhairava'. The Saiva version (8.10cd) has tasya madhye 
nyased devi bhairavam bhlmavikramam 'O Devi, in the centre of that [lotus] he 
should install Bhairava of terrible might'. But for this parallel we might have 
been tempted to read the Buddhist version not as mahdbhairava bhlsanam, with 
mahabhairava as a stem-form substituted for the accusative mahabhairavam 
for metrical convenience, a common licence in this register of the language, 
but as mahabhairavabhlsanam, preferring a pleonasm 'most frightening [and] 
terrible' to a reading that shows the name of the deity of the Vidyapitha, a clear 
sign of incomplete assimilation. 

The Saiva text follows this with tasyagratah sthitam devim aghoram 
ghoravikramam \ bhairavabhimukhlm kruddham raudrarupam nyaset tatah 
'Then he should install the goddess Aghora of frightening might standing before 
him, facing Bhairava, furious and of terrible aspect'. The Buddhist version 
first inserts a description of some of the male deity's iconographic features 
and then returns to redact its Saiva exemplar as follows: tasyagratah sthitam 
devim vajravarahlm sughoram \ mahabhairavabhimukham krtva tu trinetrlm 
raudrarupinlm '[and] the most frightening goddess Vajravarahi standing be- 
fore him, three-eyed, of terrible aspect, making her face Mahabhairava'. The 
Buddhist name of Heruka's consort has been inserted but the redactor has not 
troubled to do the same for Heruka, leaving the Saiva name unchanged. The 
accessible Sanskrit manuscript does give the name of Heruka here, reading 
srlherukabhimukham krtva tu, and this reading is supported by the commenta- 
tors Bhavabhatta (srlherukabhimukham krtva) and Indrabhuti (he ru ka dpal 
ye shes phyogs ni [*srlherukajnanabhimukha-]), and the Tibetan translation (he 
ru ka dpal che la phyogs [*mahasrlherukabhimukha-]). But it is certain that 
this is a later improvement, because mahabhairavabhimukham krtva tu is what 
we find in the older redaction attested in Jayabhadra's commentary, and in the 
text as incorporated in the Abhidhanottara (mahabhairavabhimukhlm). It is 
also supported by the commentary on the later form of the Laghusamvara by 
Viravajra, who gives rab 'jigs byed che la phyogs 'facing Mahabhairava' here. 


Genesis and Development of Tantrism 

Further, in most places where a Buddhist imprint is visible the text be- 
comes unmetrical. This is most economically explained by the hypothesis stated 
above 477 that what we are seeing is a Saiva source after its redaction by a Bud- 
dhist with little concern for metrical accuracy 478 

Finally the Laghusamvara's account of initiation is remarkably un- 
Buddhist in its content. This is not so much because it adheres so closely to the 
structure and detail of the ceremony outlined in the Yoginisamcara, including 
such distinctive details as the pitching of the lines of the Mandala with a cord 
soaked with human blood and made from the hair or sinews of a corpse (2.11), 
the use of such substances as the five nectars of the body (pancamrtam) and the 
ash and powdered charcoal of cremation pyres on the ground of the Mandala 
(2.1-3), 479 the beating of a drum in its worship (3.2-3), and the marking of the 

477 See here p. 190. 

478 See 2.4ab: samyagjhdnatantrajhah srlherukamantrajhah; 2.13d: yajed 
ddkinljdlasamvaram; and 3.7a: sriherukddipitha darsayet. The reading 
mahdbhairavdbhimukhdm krtvd tu (2.18a) probably represents a first attempt to 
differentiate the Buddhist version from its metrical Saiva prototype by adding 

479 This substitution of inauspicious and dangerous substances in the preparation of 
the Mandala is a marked feature of accounts of initiation found in Vidyapitha 
texts. See, e.g., Picumata f. 5vl (3.12ab), concerning the Aghorimandala): asthi- 
curnatadahgdraih mantrajno dlikhet puram 'The mantra-master should draw the 
Mandala with powdered bone and charred bone'; f. 5v6 (3.31ab): smasdnotthena 
sutrena sutrakdryam tu kdrayet 'He should do the outlining with a cord from 
the cremation ground'; f. 10r2-3 (3.184—185): smasdnotthdni bhdnddni vas- 
trasutrddikdni tu \ vastrai<r> dhvajd tu kartavyd sutrena karani tathd \\ kesair 
darbhd<n> yathdnydyam *acchinndgrdn (corr. : acchinndgrdh Cod.) prakalpayet 

| vestayen mandalam tais tu astrajaptaih samantatah 'The vases, cloths and cords 
should be made with what has come from cremation grounds. With [funerall 
shrouds he should make the banners and with threads [therefroml the pitch- 
ing cord. With the hair [of corpses] he should provide the uncut-ended stems of 
[protective] Darbha grass. After empowering them with the weapon[-mantra] he 
should surround the mandala with them'; Jayadrathaydmala, Satka 4, f. 65v7 
(Rdviniydgapatala, [concerning the Mandala of Ravini in the Kdlikula section of 
the Jayadrathaydmalatantra], v. lOlcd: savasutrena samsutrya asthicurnddibhir 
likhet 'He should colour [the Mandala] with powdered [human] bone and the like 
after pitching its lines with a corpse-cord'; Jayadrathaydmala, Satka 3, f. 200r5- 
6: sutrayed rudhirdktena *savasutrena (corr. : savasutrena Cod.) 'He should 
outline the Mandala with a corpse-cord smeared with blood'. The nature of 
this cord is indicated by Ksemaraja on Svacchandatantra 13.21b: mrtasutrena 
vaksyamdnacchummakdyuktyd mrtasndyund 'The expression 'with a corpse-thread' 
means 'with the sinew of a corpse' in accordance with the secret vocabulary to 
be taught below'. He refers here to Svacchandatantra 15.5: sndyuh sutram 
prakirtitam 'The word cord means sinew'. This understanding is also seen in Bud- 
dhist Tantric literature. In his commentary (-pinddrthatikd) on the Hevajratantra 
Vajragarbha glosses smasdnasutrena 'cremation ground cord' as ro'i rgyuspa mams 
kyis byas pa'i srang bus 'a cord made from the sinews of a [human] corpse' (Snell- 
grove 1959, Pt. 1, p. 51, n.l, who mistranslates this to refer to 'a thread made 


The Saiva Age 

foreheads of the candidates with human blood (3.8). 480 It is more because the 
redactor has not added what from the time of the Mahavairocanabhisambodhi 
onwards had been the most marked characteristic of the Mantranaya's adap- 
tation of Saiva Mandala initiation, namely the series of consecrations known 
as abhisekah. The commentators evidently could not accept that this crucial 
Buddhist signature might be absent. For they have resorted to strained exegesis 
in order to impose it. Jayabhadra claims that the terse injunction to worship 
the Mudra in 3.7 alludes to the guhyabhisekah, in which the Guru unites with 
a consort (mudra) and the candidate swallows the semen. Then avoiding the 
difficult task of reading in allusions to any of the six consecrations that normally 
preceded this climactic act in his time he simply asserts that they should be 
done following the procedure familiar from other Tantras. 481 Bhavabhatta, 
however, adopts a more bold and imaginative strategy, finding all seven con- 

front the guts of a corpse')- We also read of the use of the hair of corpses for 
this purpose: Jayadrathayamala, Satka 3, f. 181r4: dlikhen mandalavaram tato 
raudrena bhasmand | prathamam sutrayitvd tu savamurdhajarajjund 'He should 
draw the excellent Mandala with human ash after first pitching its lines with a cord 
of corpse-hair'; Siddhayogesvarlmata 8.8: narakesasamutthena karpdsddimayena 
vd | sutrayen mandalam divyam sarvasiddhiphalodayam 'He should trace the ex- 
cellent Mandala, which bestows the reward of all the Siddhis, with [a cordl made 
from human hair or from fibres such as cotton'. This option is no doubt fixed: 
cremation-ground substances for ascetics and conventional substances for house- 
holders; see, e.g., Jayadrathayamala, Satka 2 f. 9v2 (VdmesvarTydgapatala, w. 
48c— 49): vdmdmrtddibhir lipya tatra mandalam dlikhet || rajobhir *viramdrgasthas 
(em. : viramdrgasthais Cod.) *cityahgdrddibhasmabhih (cityahgdrddi conj. : 
citydhgdrddi Cod flc : citdhgdrddi Cod'") | ratnddisdlijdtais ca grhasthas cdlikhet 
tatah 'Having smeared [the ground] with wine and the like he should draw the 
Mandala upon it with powders such as the charcoal and ash of funeral pyres, if he 
follows the path of Heroes, and with [ground] precious stones or rice flour [etc.], if 
he is a housholder'. 

480 Both versions say only that this is to be done 'with blood' (raktena). But a variant 
specifying human blood (mahdraktena) is attested by the Tibetan translation (mt- 
shal chen gsum Ian bzlas pa yis [mahdraktena trijaptena]) and the commentators 
Durjayacandra (mtshal chen Ian gsum brzlaspayis), Viravajra (id.), and Indrabhuti 
(mtshal chen ni). 

481 Jayabhadra, Cakrasamvarapanjikd, p. 114, 11. 9-11: kulam tasya vinirdised (3.6) 
itiparyantam sukaram eva | tadanantaram tantrdnantaraprasiddhena vidhind sar- 
vam abhisekam nivartyeddnlm guhydbhisekavidhipradhdnatvdt pujayen mudrdm 
(3.7) ityddind guhydbhisekam sucayati 'The text up to 'he should indicate his Fam- 
ily' is easy. He now alludes to the guhyabhisekah with the words beginning 'he 
should worship the Mudra'. He does so because this is the most important [of the 
consecrations]. [It is should be understood that] 'he should worship the Mudra [i.e. 
the consort] after he has completed the whole consecration [process that should be 
performed] immediately after that [determining of the candidate's Family by cast- 
ing the flower] following the procedure that is well known from other Tantras'. The 
expression 'the whole consecration', though singular, should be understood to refer 
to the whole sequence of the consecrations that precede the guhyabhisekah. 


Genesis and Development of Tantrism 

secrations up to and including the guhyabhisekah in 3. 2-3. 3a.: ghantanadam 
dlambya puspadhupair alamkrtdm \ ghantdm vddayet susvaram patahikdm 
vdpi sddhakah \ hdhdkdram ca kdrayet 'Resorting to the resonance of the bell 
the Sadhaka should ring the bell after it has been adorned with flowers and 
[fumigated with] incense; or he may [beat] a drum. He should also laugh wildly'. 
He asks us to accept that the ringing of the bell refers to the consecration of 
[the giving of] the bell (ghantdbhisekah) 482 and, more astonishingly, that the 
wild laughter enjoined, literally 'the sound hd ha', is the consecration of [the 
giving of the initiatory] name (ndmdbhisekah). 483 Having conjured up these two 
consecrations he then asserts that the three that precede them are therefore 
implicitly intended, namely the consecration with water (udakdbhisekah), the 
consecration with the crown (makutdbhisekah), and the consecration with the 
Vajra (vajradhipatyabhisekah). 484 He then subjects this same passage to a 
second reading in order to force it to refer also to the two consecrations that 
follow these five: the deary abhisekah, which qualifies the initiate to officiate as a 
Vajracarya, and the consecration of the secret (guhyabhisekah). He claims that 
in this second reading the resonance of the bell, the ringing of the bell, and the 
beating of the drum refer to the Guru's uniting for the purpose of the second of 
these consecrations with a girl of twenty-five, twelve, or sixteen respectively 485 

482 Bhavabhatta, Cakrasamvarapanjika, p. 37, 1. 17: ghantanadam ityadina ghantd- 
bhisekah pratipddyate 'The passage beginning with ghantanadam teaches the con- 
secration of the bell'. 

483 Bhavabhatta, Cakrasamvarapanjika p. 38, 11. 6-7: hdhdkdram ca kdrayed iti | 
hdhdkdro ndmdbhisekah \ tarn gurubhattdrakendtmanah kdrayet 'In the expression 
"He should have the hdhdkdrah done", the hdhdkdrah is the consecration of the 
name. He should have that done for himself by the venerable Guru'. 

484 Bhavabhatta, Cakrasamvarapanjika, p. 38, 1. 10: tata udakamaulivajrddhi- 
patyabhisekdndm grahanam tatpurvakatvdt tayoh 'From this [reference to the con- 
secrations of the bell and the name] it follows that the text also refers [by implica- 
tion] to the consecrations of water, crown, and the Vajra Lord, because those two 
have to be preceded by these [three]'. The five consecrations covered here are as in 
Samvarodaya 18.27, where they are associated with the five Tathagatas. 

485 Bhavabhatta, Cakrasamvarapanjika, p. 38, 11. 13-14: ghantdnddah *svalihgd- 
vasthitapahcavimsatikddhidhdnam (em. : svalihgdvasthitah pancavimsatikdbhi- 
dhdnam Ed.) | ghantd dvddasdbdikd | patahikd sodasdbdikd | ghantdnddo vajra- 
kulam | ghantd ratnakulam | patahikd padmakulam | hdhdkdras tathdgatakulam | 
cakdrdd anyac ca | * ghantdnddddlndm anyatamdm dedryah sevayet (em. : ghantd- 
dindm anyatamdnoedsevayet Ed.) | ghantanadam aho sukheti mantram sddhakah 
sisyah kdrayed uccdrayed ity arthah | kuto 'nantaram ity aha | andmetyddi | andmd- 
hgusthavaktrdbhydm lehayed yogavit sadd | somapdnavad dsvddya siddhim dpnoti 
sdsvatlm (1.12c-13a) iti gdtheha yojitavyd | tato 'syd idam arthdntaram \purvokta- 
prajhdsevayd yad bhutam tad andmdhgusthavaktrdbhydm dedryah sisyam lehayet 
| sa ca sisyah tatah somapdnavad dsvddya siddhim dpnotlti guhydbhiseko 'yam 'The 
resonance of the bell denotes a girl of twenty-five mounted on one's penis; the bell is 
a girl of twelve; and the drum is a girl of sixteen. [In addition] the resonance of the 


The Saiva Age 

Having made the text refer to the guhyabhisekah, he finds the acaryabhisekah 
by using the same argument that he had employed to arrive at the full sequence 
of the five consecrations that precede it, namely that its presence is entailed 
by the supposed reference to the guhyabhisekah, because that requires it as its 
antecedent. 486 He finds a reference to the final consecration that he needed to 
discover here, that of wisdom (prajnabhisekah), in the statement in 3.7 that 
Jayabhadra had taken to allude to the preceding guhyabhisekah: tatah pujayen 
mudram acaryah susamahitah 'Then the Acarya, fully concentrated, should 
worship the Mudra'. If, as is highly probable, the consecration understood by 
Bhavabhatta here was the prajnajnanabhisekah of the initiation manuals, then 
there would appear to a problem, because the active agent in that consecration 
was not the Acarya but the candidate, who now unites with the consort himself. 
Bhavabhatta is very terse at this point but it is likely that he was attempting to 
remove this difficulty when he wrote that the text refers to the agent as Acarya 
here because he is endowed with such qualities as self-control. I take him to 
mean that it is indeed the candidate rather than the officiant that is the agent 
here and that he is referred as an officiant only figuratively, because he has all 
the qualities that are required of an officiant. 487 These readings are, of course, 

bell is [a women of] the Vajra Family, the bell [one of] the Jewel Family, the drum 
[one of] the Lotus Family, and the wild laughter [one of] the Tathagata Family. The 
word 'and' [in 'and he should laugh wildly' indicates [one of] the other [Family, that 
of Action]. The officiant should have intercourse with one or other of these women of 
whom the first is 'the resonance of the bell'. The meaning is [also] that the Sadhaka, 
[that is to say,] the candidate, should make, that is to say, utter, 'the resonance of 
the bell', that is to say, the Mantra aho sukha ['Oh, Bliss']. He [also] tells us that 
after which [he should utter this Mantra] in the passage [of this Tantra] that begins 
with anama-. At this point one must read in the following verse (1.12c-13a) 'The 
master of Yoga should always lick [it, taking it] with the tips of his ring finger and 
thumb. Having relished it as though it were a draught of Soma he attains eternal 
success'. So there is another sense of this [verse], namely that the officiant should 
make the candidate take into his mouth [lit. 'lick'] the product of his sexual union 
with the aforesaid consort with the tips of his ring finger and thumb; and that can- 
didate, having relished it like a draught of Soma attains Siddhi. This, then, is the 
guhyabhisekah' . 

486 Bhavabhatta, Cakrasamvarapanjika, p. 38, 11. 23-24: sa ca sisyah tatah 
somapanavad asvadya siddhim apnotlti guhyabhiseko 'yam | ata ev acaryabhisekah 
siddhah tatpurvakatvat tasya 'This is the guhyabhisekah. This itself establishes 
the presence of the acaryabhisekah, because the former is preceded by the latter'. 

487 Bhavabhatta, Cakrasamvarapanjika, p. 39, 11. 21-22: tata ityadina prajnabhisekam 
darsayati | tato guhyabhisekanantaram | acarya iti dhairyadigunayogat 'In the pas- 
sage beginning tatah he reveals the Wisdom Consecration. The word tatah ('next') 
means directly after the guhyabhisekah. He is termed the officiant [here] because 
he has such qualities as self-control'. Bhavabhatta is probably alluding to the qual- 
ities of the good Acarya as stated in vv. 8-9 of the Gurupancasika: dhiro vinlto 
matiman ksamavan arjavo 'sathah | . . . . 


Genesis and Development of Tantrism 

artificial and could be imposed on the text only because Bhavabhatta, like 
Jayabhadra, could not accept the possibility that there might be no reference to 
the consecrations in a Buddhist Tantra's treatment of initiation. 

Further exemplification of the direction of redaction can be seen in the first 
of the new parallels listed above, that on the subject of the regular rite of wor- 
shipping the Kulika (as the Laghusamvara has it). For ease of comparison I give 
in bold characters those parts of each of the three related texts, the Picumata, 
the Herukabhyudaya, and the Laghusamvara, that partly or completely corre- 
spond to passages in one or both of the other two. The Picumata passage is as 

mulasutrddikdndm tu kramam sadhanalaksanam \\ 

10 durlabham trisu lokesu samayacarapalanam \ 
ydgam vidhis tatha jnanam cakram yogam ca sobhanam || 

11 kathaydmi mahadevi yat tvayd coditam *balam (?) | 
madhyamottamacchagena gandhodasahitena tu || 

12 vatikam prasayet prajnah pujakale visesatah \ 
vidhdnan tu sadd yojyam carvdhdrena suvrate \\ 

13 samaye sddhane caiva dravydlabhanakarmani | 
tasyaiva dutayah siddhah sahaja viravandite \\ 

14 gurunddivibhdgena srstidravyddisamgrahe | 
rtuyogaviyogena anulomavilomajd || 

15 ydgddhordhvagatd devi sarvakdmavilaksand | 
kundagolodbhavenaiva svayambhukusumena ca || 

16 japahomarcanam sndnam bukapuspasamanvitam 
niyojyam svena mdrgena svakdle ydgapurvakam \\ 

f. 319v3-5 

lie madhyamottamacchagena em. : adhamottamacchdgena Cod. 488 

The related passage in the Herukabhyudaya is accessible only in its Tibetan 
translation. I give that here with a reconstruction of the Sanskrit of the parts 

488 j propose this emendation for two reasons. The first is that the reading contradicts 
information given later in this chapter. According to that there are three grades 
of flesh for use in the preparation of the sacrament (caruh): goat, cow, and human. 
The first is said to be inferior (adhama-), the second intermediate (madhyama-), and 
the third superior (uttama-): adhamam cchdgam ity uktam madhyamam gobhavam 
bhavet \ purusottamam mahadevi tridhd tu caravah smrtdh (f. 320r5— vl [84.36c— 
37b]). Consequently without this emendation we have nonsense: 'with the inferior 
[i.e. goat], the superior [i.e. human], and goat'. With it we have a statement that 
is consistent with this classification: 'with the intermediate [i.e. cow], the superior 
[i.e. human] and [the inferior, i.e.] goat. The second reason is that the emendation 
has the support of the Buddhist parallels, which, as we shall see, read madhyamot- 
tomasvdsena or madhyamottamocchvdsena here. 


The Saiva Age 

that match the passage in the Picumata: 

(15.6) sngags dang phyag rgya sbyar bar bya | 

dam tshig thams cad bskyang bya ste | 

jig rten gsum na rnyed dka' ba (durlabham trisu lokesu) | 

g.yon nas skyes pas byed pa yin | 

(7) dam tshig spyod pa'i mtshan nyid dang (samayacaralaksanam) \ 
sbyor nyid cho ga'i yi ge shes ni (yoga eva vidhijnanam) \ 

de ni nga yis bshad kyis nyon (tan me nigaditam srnu) \ 
dbugs dbyung mchog gi bar dag ni (madhyamottamasvasena) \ 

(8) dri yi chu dang bcas pa dang (gandhodakasahitena [tu]) | 
rtag tu ril bu bza' par bya (vatikam prasayen nityam) 
mchod pa'i dus kyi bye brag la (pujakalavisesatah) | 

pho nyas lhan cig skyes dngos grub pa (dutayah sahajah siddha) | 

(9) dman pa mchog dang 'bring mams kyi (adhamottamamadhyamah) \ 
de yis sbyor bas dngos grub 'gyur (tabhir yogena siddhih syat) \ 

'dod pa'i don kun sgrub pa'o (sarvakamarthasadhakah) | 
dpal Idan he ru ka las byung (sriherukodbhavam) \ 

(10) rang byung me tog nyid dag gis (svayambhukusumair api) \ 
cho ga shes pas kun tu spyod (vidhijhanasamacara-) | 

bzlas dang bsam gtan mchod pa dang (japadhyanapuja) | 
me tog gcig dang yang dag Idan (ekapuspasamanvitam) 

Khrag 'thung mngon par 'byung ba D f. 12r6— v2 (Herukdbhyudaya 15.6-10) 

Testimonium — Kumaracandra, Katipayaksard nama Herukabhyudayapanjika, 
p. 156: evam maya nigaditam. srnu \ madhyamottamasvasah panca pradipah 
| gandhodakam pancdmrtdni \ vatikam prasya (Cod. [f. 3v6] : prapya Ed.) 
*bhavanaganamandaladau (bhdvandgana corr. : bhavand gana Ed.) dutim 
pujayet | adhamah mantrajah | uttamah sahajah | madhyamah ksetrajah | 
tdbhih siddhih syat tasya yoginah. 

The version of the Laghusamvara reads: 

1.4 sambhavan nadarupad viniskrantah samayacaragocarah \ 
durlabham trisu lokesu adimadhyantasamsthitam \\ 

5 manthyamanthanasamyogam yatha tatha mantrajapadhydnadiyuktam | 
yogas caiva vidhijnanam tantre nigaditam srnu \\ 

6 madhyamottamocchvasena gandhodakasahitena tu 
kulikam pujayen nityam kalavisesena tu || 

7 dutayah sahajah siddha adhamottamamadhyamah \ 

f. lv2-5 

6a madhyamottamocchvasena JAYABHADRA : madhyamottamasvasena Cod. 

The Herukdbhyudaya, then, shows a version that is closer than the 


Genesis and Development of Tantrism 

Laghusamvara to the text of the Picumata in some details and covers more of 
it. It is particularly striking that it preserves the Picumata's vatikam prasayet 
prajnah pujakale visesatah (84.12ab), reading rtag tu ril bu bza' par by a \ 
mchod pa'i dus kyi bye brag la 'Let him always swallow the sacramental pellet, 
especially at the time of worship', diverging from the Picumata only in having 
nityam (rtag tu) and pujakalavisesatah where that has prajnah and pujakale 
visesatah . That the Sanskrit read vatikam is shown by the gloss vatikam prasya 
in the Herukabhyudayapahjika (f. 3v6). 

Even so it shows signs of having had difficulty in understanding some of 
the Saiva proto-text's technical terms and of having dealt with this difficulty by 
resorting to rewriting. Thus in 15.10 me tog gcig dang yang dag Idan 'together 
with a single flower' corresponds to bukapuspasamanvitam 'together with the 
Buka flower' in Picumata 84.16, so that the Sanskrit may be restored from 
the Tibetan with some confidence as ekapuspasamanvitam. The context is a 
listing of impure ingredients to be consumed at the time of practice. Now, 'a 
single flower' yields no appropriate sense in this context, whereas 'Buka flower' 
{bukapuspam) does. For the Picumata tells us that in its secret vocabulary 
bukam means 'the impurity of the male organ' (84.38a: buko lingamalo jneyas; 
87.196d: bukam lingamalam smrtam), and the Kubjikamata tells us that 
bukapuspam has the same meaning (25.226ab: bukapuspa kanakhyam ca 
lingapankamalam tatha). It is probable that the Buddhist redactor, failing to 
understand this obscure term, modified the text to produce something that had 
at least the appearance of sense. Kumaracandra confirms the reading ekapuspa- 
in his Herukabhyudayapanjika and ventures to explain it as 'the blood of a 
[woman's] first menstruation': ekapuspam prathamam rajah vajrapadmabhyam 
sadhyamanam kapalastham (p. 156) '[After putting it] in a skull-bowl [he should 
swallow] the 'one flower', i.e. the first menses, produced by the penis and vagina'. 
But this gloss is not only strained: it also leads the text into an implausible 
repetition, since the blood of first menstruation has just been mentioned in 
15.10a, in the term rang byung me tog (= svayambhukusumam). He also seems 
not to have understood the expression kundagolodbhava- seen in Picumata 
84.15c (kundagolodbhavenaiva), another 'secret' Vidyapitha term, referring to 
the mingled ejaculates. He resolves his quandary by substituting the name of 
his deity, the Tibetan dpal Idan he ru ka las byung (15. 9d) evidently rendering 

In the abbreviated version seen in the Laghusamvara we have kulikam 
pujayen nityam 'let him constantly worship the Kulika' in place of the reading 
vatikam prasayen nityam seen in the Herukabhyudaya and in the Saiva proto- 
text. This is evidently the result of a corruption of a redaction which read not 


The Saiva Age 

vatikam but the exact synonym gulikam; 489 and this hypothesis is confirmed by 
the Abhidhdnottara, which in its own first chapter preserves gulikam in a pas- 
sage modelled on these verses of the Laghusamvara, thus bearing witness to a 
stage of the redaction of this text that is earlier even than that known to our ear- 
liest commentator, since Jayabhadra accords with all later witnesses in reading 
kulikdm here: 

1.7 nadarupad viniskranta samayacaragocaram 
durlabham trisu lokesu ddimadhydntanirmalam || 

8 manthamanthdnayogena samyogad yatra yat tathd | 
prakrtiprabhasvaram suddham guhyapithodbhavodbhavam || 

9 nirdosam sasvatam sdntam khasamam srstikdrakam | 
svabhdvasuddham svayambhutam yoginindm sukhapradam \\ 

10 jdpadhydnddibhir yuktam yogasyaiva vidhijnatd | 
tantre nigaditam tattvam guhyakddhipate srnu || 

11 madhyamottamasvasena gandhodakasahitena tu | 
gulikam karayed dhiman pujayet parsamandalam 490 || 

12 kalavelavisesena pujayet tatra dutayah | 
sahajah siddhidah sarvd adhamottamamadhyamdh || 

13 antargatena manasd kamasiddhim tu sddhayet 

Abhidhdnottara A f. 2r2-6; B f. 2r4-v3 

7ab nadarupad em. : nddarupo B : nddaru + A • viniskranta samayacaragocaram 
B : + + + + + + cdragocaram A 7d nirmalam A : nirmmalah B 8b samyogad yatra 
tatra yathd B : sayogdd yatra yat tathd A : *yatra tatra yathd tathd (Tib. srub dang 
bsrub par yang dag sbyor | gang la de la ji Itar bzhin) 8cd prakrtiprabhasvaram 
suddham guhyapithodbhavodbhavam B : prakrtiprabhdsva + + + + + + 
thodbhavodbhavam A 9a sasvatam A : sdsanam B 9c suddham svayambhutam 
conj. [= Tib. dag pa rang byung ste] : suddhasambhutam B : suddham adbhutam 
A lOab dhydnddibhir yuktam yogasyaiva B : dhydnddibhir yu + + + + va A • 
jhatd A :jneyd B 

The otherwise unattested kulikdm was then construed by force to mean yoginlm 
'a/the Yoginf, and the verb prasayet 'let him swallow', since it now made no sense, 
altered to pujayet 'let him worship'. 491 

489 For gulikd (variant forms: gutikd andgudikd) see here p. 217. 

490 The reading of lie is further supported by the Tibetan translation: mkhas pas dril 
bur byas nas ni. Note that dhimdn (mkhas pas) here is synonymous with prdjnah 
found at the corresponding point in the version seen in the Picumata (vatikam 
prasayet prdjnah). This, then, has probably survived from the Saiva source on 
which the first Buddhist version drew. 

491 Bhavabhatta, Cakrasamvarapanjikd, p. 20: kulikayogini | tarn pujayed drddhayet 
| nityarn sarvakdlam pratidinam ity arthah '[The word] kulikd [means] yogini. It 
is she that he should propitiate [in this way]; and he should do so constantly, at 
all times, that is to say, every day'. Cf Jayabhadra, Cakrasamvarapanjikd, p. 110: 
kulikdm iti tantre samayabhdsd \ vajravdrdhisvarupdm bdhydngandm pujayed 


Genesis and Development of Tantrism 

That the Buddhist versions arose from Saiva prototypes is clear from the 
detailed analysis of these and many other parallels. Other features reinforce 
this conclusion. In all cases the Saiva passages fit neatly into the contexts in 
which they occur, without ragged edges, as it were, at their beginning and end, 
whereas this is often not so with the parallels in the Buddhist texts, a circum- 
stance that fits well with a scenario in which the latter where constructed by a 
rather careless process of extraction, insertion, and superficial editing. 

The same is suggested by the high degree of divergence between the various 
Buddhist commentators in their attempts to tell us what these new texts mean. 
They were caught out, as it were, by new materials that lacked roots in the 
Buddhist textual corpus in which they were trained. They did their best to make 
sense of what were in many cases barely intelligible passages; but without much 
guidance from existing Buddhist sources and with no central authority to impose 
consistency on their efforts they were bound to diverge. 

We have a good example of this in the passage just discussed, in the words 
madhyamottamasvasena gandhodakasahitena tu . . . . The meaning of the Saiva 
prototype as seen in the version of the Picumata, namely madhyamottamaccha- 
gena gandhodasahitena tu \ vatikam prasayet prajnah, is perfectly clear to any- 
one who has read the whole chapter of which it is part. It means 'The wise 
[initiate] should swallow a pellet made from beef, human flesh, or goat mixed 
with scented water'. 492 The case is very different with the Buddhist versions. 
Their madhyamottamasvasena surely began life as a copyist's corruption; for it 
yields no sense in either Saiva or Buddhist terms in the context of this rite of 
the pellet or, indeed, in any other. Kumaracandra, therefore, in his commen- 
tary on the passage as it appears in the Herukabhyudaya, could only guess at 
the meaning on the basis of the one part of the sentence that made undoubted 
sense, namely the injunction to swallow a pellet. Knowing that such pellets were 
made in practice from the five meats and the five body nectars he tells us that 
madhyamottamasvasah 'the intermediate and upper breath' means those meats 
and that the gandhodakam 'scented water' with which this 'breath' is to be mixed 

iti | yathd samtoso jdyate tathd karanlyam ity arthah 'The word kulikam is used in 
[this] Tantra following [its own special] convention. It refers to the physical woman 
[who is the practitioner's consort, when she is perceived as] identical with Va- 
jravarahi. He should worship her, which means that he should do whatever is neces- 
sary to satisfy her'. In his Kdlacakra -influenced commentary on the Laghusamvara 
(Laghutantratikd) Vajrapani interprets kulikd more esoterically as referring to Va- 
jravarahi as the non-conceptual central energy-channel: kulikam pujayen nityam 
iti | iha kulikd madhyamdvadhuti vajravdrdhi nirdvarand grdhyagrdhakavarjitd 
(p. 59). 
492 See the footnote on my emendation madhyamottamacchdgena on p. 212. 


The Saiva Age 

means those nectars. 493 

Jayabhadra and Bhavabhatta commenting on the same expression when it 
occurs in the Laghusamvara, where the second part of the sentence has emerged 
through further confusion as kulikam pujayet, impose quite different but equally 
arbitrary interpretations, which are based not on the text itself but, in the ab- 
sence of evident meaning, on their own notions of what the text ought to be 
saying here. Thus Jayabhadra, who has the variant madhyamottamocchvasena, 
makes madhyama- mean 'vagina', uttamocchvasah 'the placing of the tongue', 
and gandhodakam 'semen', interpreting the sentence to mean that the adept 
should worship the Kulika, that is to say, his female consort identified with 
Vajravarahi, by placing his tongue (uttamocchvasena) together with his semen 
(gandhodakasahitena) in her vagina {madhyama-). , 494 

493 Kumaracandra, Herukabhyudayapahjika, p. 156: madhyamottamasvasah pahca 
pradlpah | gandhodakam pahcamrtani 'The word madhyamottamasvasah means 
the five 'lights'; and gandhodakam means the five nectars'. On the five lights 
and five nectars see, e.g., Vagisvaraklrti, Tattvaratndvalokavivarana 18: panca- 
pradipasabdena gokudahanalaksanasya amrtasabdena vimumarasulaksanasya 
satatanusthanam eva sadhyam manyante '[The learned] hold that the expres- 
sion pancapradipa- refers to the accomplished regular practice of the [five meats] 
of the cow (go-), dog (kufkkura]-), horse (dafmya]-), elephant (hafsti]-), and 
man (na[ra]-), and the expression amrta- to that of excrement (vi[t]-), urine 
(mu[tra]-), flesh (mafmsa]-), blood (ra[kta]-), and semen (su[kra]-)\ Cf. Jaya- 
bhadra, Cakrasamvarapanjika, p. 108: adau tavan manonukule sthane nisadya 
pancamrtakrtagulikam mukhe krtva . . . 'At the beginning [before he begins the 
Sadhana] he should sit in a place conducive to meditation, place a pellet of the 
five nectars in his mouth, ...'; Bhavabhatta, Cakrasamvarapanjika, p. 24: goku- 
dahananam pancdmrtasya ca vatikam bhavanarambhe bhaksayet At the begin- 
ning of his meditation he should swallow a pellet consisting of [the flesh of] cow, 
dog, horse, elephant, and man, and the five nectars'; Sadhanamala 251 (Advaya- 
vajra, Saptaksarasadhana), p. 490: yogi pratar utthaya samayagulikam mukhe 
praksipya . . . 'The meditator, having risen before sunrise and placed a Samaya 
pellet in his mouth . . . '. The term samaya- in samayagulika means the five nectars; 
see Bhavabhatta, Cakrasamvarapanjika p. 18: samayapalanam samayaraksanam 
pancamrtabhaksanam 'maintaining the samaya- means keeping the pledges [and] 
swallowing the five nectars'; Jayabhadra, Cakrasamvarapanjika, p. 109: samayo 
dvividhah raksaniyo bhaksaniyas ca 'The samayah is of two kinds: that which is to 
be maintained [i.e. the post-initiatory pledges] and that which is to be swallowed 
[i.e. the five nectars]'. 

494 Jayabhadra, Cakrasamvarapanjika, p. 110 : madhye bhavatiti madhyamah | 
padma ucyate | tasminn uttamocchvaso jihvavinyasah \ tena kimbhutena | gandho- 
dakasahitena tu bodhicittasahitenaivety arthah | kulikam iti tantre samayabhasa 
| vajravarahlsvarupam bahyahganam pujayed iti \ yatha samtoso jayate tatha 
karaniyam ity arthah 'The word madhyama- , meaning 'that which is in the centre', 
refers to the Lotus [i.e. the vagina]. The word uttamocchvasah means 'the placing 
of the tongue' [and madhyamottamocchvasena is a locative Tatpurusa compound 
meaning 'by the placing of (his) tongue] in that. The words gandhodakasahitena tu 
'together with the scented water' describe that [placing of his tongue in her vagina] 
and mean that it should be together with [his] Intention to Attain Enlightenement 


Genesis and Development of Tantrism 

In Bhavabhatta's commentary we find an entirely different understanding. 
According to him madhyamottamasvdsena gandhodakasahitena tu \ kulikam 
pujayet means 'he should worship the Yogini with the place or time (-svasena) 
of fire (madhyama-) and earth (-uttama-) together with wind (gandha-) and 
water {udaka-)\ The purpose of this invention, which the Sanskrit entirely 
fails to support, is to find a reference (1) to the symbols of the four elements as 
constituting the thrones of the various groups of Yoginis in the Mandala and (2) 
to various time periods considered to be governed by these elements as the occa- 
sions for the successful performance of rituals for hostile purposes (abhicarah), 
re-invigoration (paustikam), expulsion (uccdtanam), and the averting of danger 
(santikam) respectively That Bhavabhatta has decided what he would like to 
find here and then imposed it is clear from the extreme artificiality of the glosses 
that bend the text to his will: 'the intermediate' (madhyama-) is fire (vahnih) 
because it is falls in the middle of the list of the four elements (actually in the 
penultimate position); the 'highest' (-uttama-) is that of Mahendra, the presiding 
deity of the symbol of earth (prthivl), because he is the king of the gods; gandhah 
means not 'fragrance', its lexical meaning, but 'that which possesses fragrance', 
namely the wind (vayuh), since that is the bearer of fragrance; udaka- is not 
udakam 'water' but an unattested udakah meaning Varuna, literally 'he who 
possesses the waters', since Varuna is the presiding deity of the symbol of water 
(udakam); and svasah means not 'breath' but 'that in which X breathes', that is 
to say, by an entirely unwarranted leap, the locus or time of X's operation. 495 

[i.e. his semen]. The word kulika is a term specific to the esoteric jargon of this 
Tantra. It denotes the physical woman [as] identical with Vajravarahi. By saying 
that one should 'worship' her the text means that one must do what is necessary to 
satisfy her'. 
495 Bhavabhadra, Cakrasamvarapahjika , p. 20: madhyama uttamah svasity asminn 
aneneti va \ svasah sthanam kalo va \ madhyamo vahnih prthivyaptejovayava iti 
vacanena madhyodbhavatvat | \madhyodbhutatve py upayagrahanam yatas tarn 
vaksyati] | uttamo mahendro devarajatvat | madhyamottamayoh svasah sthanam 
kalo vety arthah | tena kulikam pujayed iti sambandhah | kimbhutenety aha gandhe- 
tyadi \ gandho 'syastiti gandho gandhavahatvad vayuh \ udakam asyastity udako 
varunah | tayoh sthanena sahito gandhodakasahita iti madhyapadalopT samasah 
ghrtapurno ghato ghrtaghato yathd 'The term svasah is to be understood here to be 
derived from the root svas 'to breath' in the meaning 'that in which X breathes', X in 
this case being madhyamah and uttamah. The svasah, then, is the locus of these or 
their time-period. The madhyamah 'intermediate' is 'fire', because it arises in the 
middle, in accordance with the text 'earth, water, fire, and wind'; and the uttamah 
'highest' is the [symbol] of Mahendraf, the presiding deity of the earth symbol], be- 
cause he is the king of the gods. So the meaning of madhyamottamasvasah is 'the 
locus or time of the madhyamah and the uttamah'. With this he should worship 
the Kulika. Such is the core syntax. The compound beginning gandha- describes 
this svasah further as 'accompanied by gandha- and udaka-', meaning 'together 
with the locus of these [other] two (gandhodakasthanasahitah). This is a com- 


The Saiva Age 

Since these confused and barely comprehensible verses are found in the 
opening chapter of the Laghusamvara the redactor has made a greater effort 
than usual to assimilate them to their new Buddhist milieu. But he has not 
done this by rewriting them in such a way that Buddhists would recognize and 
understand them as formulated within their own established discourse. His ap- 
proach is rather that of montage or bricolage, in which bits and pieces of various 
texts have been clumsily combined. Instead of rewriting the verses he has sand- 
wiched them between others derived from well-known Buddhist sources. Thus 
the opening verses of the work (1.1-3), which immediately precede this pas- 
sage, are a version of the opening of the Buddhist Sarvabuddhasamayoga; 496 
and the verses (1.7c-13b) that follow it contain awkwardly collocated variants of 
verses found in that text and the Buddhist Sarvatathdgatatattvasamgraha and 
Guhyasamaja. 497 

But this attempt to lend the compilation a Buddhist character by embedding 
the passage from the Saiva Vidyapitha between verses that Buddhist Tantrics 
would immediately recognize as Buddhist is mostly restricted to this first sec- 
tion. The rest of the work up to the point at which the redaction known to Jaya- 

pound of the type in which an intermediate word is dropped, as when one says 'a 
pot of ghee' (ghrtaghatah) when what one means is 'a pot full of ghee'. The other 
two are gandhah and udakah. The first is a primary derivative of gandhah 'fra- 
grance' in the meaning 'that which has fragrance' and refers to the wind, because 
that is the bearer of fragrance. The second is [likewise] a primary derivative of 
udakam 'water' in the meaning 'that which has water', i.e. Varunaf, the God of Wa- 
ter]'. The application of this explanation then follows. One is instructed to meditate 
on the Yoginis one by one in a fixed order of rotation tied to the passage of time. 
Thus on the first Tithi of the lunar fortnight one meditates on the first eight Yo- 
ginis during the day-time, each for one eighth of the day, the second eight during 
the eight half Praharas of the night, the third eight during the day of the second 
Tithi, the fourth eight during the night, and so on. Bhavabhatta explains there that 
the three eights that make up the 24 Yoginis associated with the sacred sites must 
have the symbols of fire, water, and earth as their thrones (pp. 21-22: devindm 
asanam vahnimandalam iti dinabhagah; devindm asanam vdrunamandalam iti 
rdtribhdgah; devindm mdhendramandalam asanam iti dinabhagah). This, evi- 
dently, is what he means by svdsah in the sense of 'place'. He explains its sec- 
ond meaning as 'time' in the following: agnyddiyogo 'py abhicdrddau tathaiva 
jheyah | yathdbhicdre cittacakrasya vahniksane sdntike vdkcakrasya varunaksane 
paustike kdyacakrasya mdhendraksane uccdtane smasdnacakrasya vdyuksane yo- 
ginindm anyatamd bhdvyd 'This application of fire and the others should also be 
understood in the case of hostile rites and the like. Thus in a hostile rite one should 
meditate on one of the Yoginis of the Circuit of Mind (the first eight) at a fire mo- 
ment, on one of those of the Circuit of Speech (the second eight) at a Varuna moment 
in a rite to avert danger, on one of those of the Circuit of the Body (the third eight) at 
a Mahendra moment in a rite of re-invigoration, and on one of those of the Circuit of 
the Cremation Grounds (the fourth eight) at a wind moment in a rite of expulsion'. 

3 See here p. 154. 

7 See here p. 163, parallels 1, 5, and 6. 


Genesis and Development of Tantrism 

bhadra and Bhavyakirti ends consists almost entirely of (1) sections for which 
I have found close Saiva parallels, (2) sections for which I have not found such 
parallels but which are of the same type, and (3) sections devoted to giving the 
Mantras. These, of course, have not been lifted directly from Saiva sources, be- 
cause the Mantras so taught are peculiar to this and related Tantras. However, 
the Mantras themselves are Saiva in style; and the method of teaching them by 
giving them letter by letter in encoded form (mantroddharah) has been adopted 
in imitation of Saiva scriptural practice, appearing first, as we have seen, in the 
Sarvakalpasamuccaya that supplements the proto-Yoginitantra Sarvabuddha- 
samayoga. 498 In the light of this one readily understands why the redactor of 
the version known to Bhavabhatta and the other later commentators and seen 
in the one accessible manuscript and the Tibetan translation felt the need to 
add explicitly Buddhist material at the end of the work, thus accomplishing for 
the whole an unambiguously Buddhist frame, which in the earlier redaction had 
been present only in the first chapter. 499 

Converting the Outsiders. The textual dependence of these Buddhist 
Yoginitantras on the scriptural corpus of the Vidyapitha would surely have been 
obvious to any learned Sakta Saiva who examined them; and there is evidence 
that it was indeed noticed. We do not find this evidence in the Tantric Saiva 
literature, since the only historical data that intrude there are the spiritual ge- 
nealogies of its teachers. For the rest it is concerned purely with what it sees 
as the timeless realities of fact and injunction, and it is interested in relations 
between its own and other traditions only to the extent that it establishes a hi- 
erarchy among these traditions by ranking their various goals along an ascent 
that culminates in its own. If awareness of this textual dependence was to find 
expression in Saiva literature then it could only be in the distorting mirror of 
mythology, where the specifics of the tensions between sects could be translated 

498 See here p. 154. 

499 rp^g S p ec i a i character of the added, 51st chapter is indicated in the spiritual biogra- 
phy (mam thar) of Tilopa ascribed to Marpa (Mar pa chos kyi bio gros). For there 
the Jnanadakini and her retinue are said to have taught it to Tilopa together with 
the oral transmission (TORRICELLI and Naga 1995, p. 12): gsungs nets rtsa rgyud 
le'u nga gcig pa bshad rgyud dang bcas pa dang snyan rgyud gnang ngo. The ex- 
tended Tantra was already current when at least some of the Vyakhyatantras were 
redacted. The Adhidhanottara contains 50.20c-51.12b. It is possible that the text 
was extended first only to this point. Parts of the 50th chapter after this point are 
seen in the Samputodbhava: 50.21-23b and 24ab > Samputodbhava 5.1.16-19b; 
and 50.25 > Samputodbhava 5.1.19cd. Verses from the remainder of the longer 
text, from 51.12c to the end, are found in the Yoginisamcara and the Samvarodaya: 
51.7ab > Yoginisamcara 17.10ab; 51.13c-16b > Yoginisamcara 17.21c-24b; 51.18-19 
> Samvarodaya 32.29c-30b; and 51.21d > Samvarodaya 32.31d. 


The Saiva Age 

into accounts of the interaction of the gods with demons and men. Thus we find 
our evidence in a variant of the famous narrative of Siva's burning of the celestial 
cities of the three demons (tripuradahanam) given in the Haracaritacintamani, 
a collection of Saiva myths for the instruction of the laity compiled in the thir- 
teenth century by the Kashmirian Sakta Saiva Jayadratha. 500 

According to that account Brhaspati, the ingenious Guru of the gods, puts 
an end to the invincibility of these demons, the reward of their devotion to Siva, 
by fooling them into abandoning the worship of that deity He composes and in- 
troduces to them various texts for the visualization of Buddhist deities in which 
Siva and other Saiva deities are portrayed as their inferiors. Then, once they 
have become used to these, he adds Mantras by adapting those of the Saiva 
Tantras and composes passages giving instruction in Tantric ritual procedures 
by cobbling together various excerpts from the same sources. Finally, he com- 
poses Buddhist treatises which supplement this Tantric corpus with reasoned 
arguments designed to undermine the demons' commitment to their rites and 
belief in God: 501 

500 Jayadratha was the brother of Jayaratha, author of the Tantrdlokaviveka, on whose 
date see Sanderson 2007a, pp. 418-419. That Jayadratha shared his brother's 
Sakta Saiva adherence, in keeping with the family's long-established tradition, is 
evident throughout his work, but particularly in the opening verses of each chap- 
ter, in which he gives a metaphysical reading of the myth that follows. Thus in 
13.1, introducing this narrative of the destruction of the three cities, whose point 
is to glorify the Kashmirian sacred site of the volcanic fire-Linga (jvdldlihgam) at 
Suyam (Svayambhu) (on which see Stein 1900, vol. 2, pp. 484-485), he equates 
the three cities with the cognizer, cognition, and the cognized differentiated in 
contracted consciousness, and the fire that destroys them with the all-inclusive 
nonduality whose emergence bestows liberation: etad vedakavedyavedanamayam 
dagdhvd purdndm trayam purnddvairahutdsanena samayan mdydmayopadravam 
\jvdldlihgatayd *sphurah (A : sphuraj Ed.) jagadanugrdhl svayambhur asau devah 
samprati bhdsatdm mama param ullasayan nirvrtim 'May that god Svayambhu 
blaze forth for me now, revealing the highest bliss, he who has favoured the world 
by manifesting himself as the fire-Linga after burning these three cities that are 
the cognizer, the cognized, and cognition, putting an end to the torment of bound 
existence with the fire of all-inclusive nonduality'. This is exactly in the conceptual 
mode of the Sakta Saiva nondualism of Kashmir. 

501 Haracaritacintamani 13.61—83: ripunam bhagavadbhaktir vijaye mulakaranam 
| sa saithilyam avapnoti kena yatnena cintyatam || 62 tatrabhyupayah prdyena 
kascit samcintito maya \ sukrasya samnidhdne tu kathamkdram pragalbhate || 
63 tesam hitam *prdpayitum (conj. : prdrthayitum Codd. Ed.) sukra eva dine 
dine \ bhagavadbhaktiddrdhydya prayatnam adhitisthati || 64 svayam yady api 
He (Codd. : ye Ed.) bhaktds tathdpy aisvaryagarvitdh | mitaprajnas ca yojyante 
helayaiva viparyaye \\ 65 ity uktavan mahendrena *prcchyate (A 7 ": prcchate Ed. 
A'"BC) sma sa kautukat | bhagavan bruhi tarn yuktim tesam lihgdrcandpaham 
|| 66 srutveti so 'bravit pasya prayah sarve 'pi sarvada | uttarottaram utkarsam 
jnatva rajyanti jantavah \\ 67 tad Isvarad rte ko 'tra sarvesam murdhani sthitah 
| svavikalpena tasyapi kascid urdhvastha ucyate || 68 evam maydmayam tesam 
varnyate svopakalpitam \ sdstram ca darsyate kimcil likhitvd nijayd dhiyd \\ 


Genesis and Development of Tantrism 

[Brhaspati:] "The root cause of the victory of our enemies is their devotion to Siva. 
We must think carefully what will cause that to fade. I have already thought in 
general terms of a means of accomplishing that. But how[, I wonder,] will it suc- 
ceed while [their Guru] Sukra is with them? For he exerts himself day after day 
to strengthen their devotion to the Lord in order to *secure (conj.) their welfare. 
[But] although they are genuinely devoted [to Siva] they are proud of their power 
and of low intelligence. It should therefore be easy to lead them astray". When he 
had said this Indra eagerly asked him to explain the stratagem that would put an 
end to their worship of the Linga. Having heard this he replied and said: "Behold. 
All persons usually assign their devotion on the basis of their understanding of an 
ascending hierarchy. Who but Siva is at the summit [of this hierarchy], surpass- 
ing all [others]? Nonetheless I shall use my imagination and tell [them] that there 
is a being above even him. In this way I shall give them false instruction of my 
own invention. I shall also use my wits to compose and show them some learned 
writing [in support of my teaching]. I shall deceitfully write visualization-texts 
of deities in relation to whom this Siva will be placed in a position of inferiority, 
and I shall tell them that these show that there is another being who is greater 
even than him, so that they may give up their worship of the Linga and so be de- 
stroyed. However, these false teachings will have no effect while Sukra is present. 

69 dhydndni devatdndm ca likhyante tdni kaitavdt \ ydsdm mahesvaro py esa 
nyagbhdvena *nivesyate (Codd. : nyavesyate Ed.) || 70 evam mahesvardd anya 
utkrsta iti kathyate | tesdm yato bhavel lihgapujdsaithilyatah *ksatih (Ed. : 
ksitih A : matih BC) || 71 sukrasya samnidhdne tu prathante na kaduktayah 
| *sa pratityopapattyd (A : sapratlpopapattyd Ed. BC) ca paramdrthavisdradah 
|| 72 ity uktavdn dhgiraso vdsavena sagauravam \ abhyarthyate sma sd yuk- 
tir akhandd kathyatdm iti || 73 uvdca sa tatah *sakram (A : sukram Ed. BC) 
dkalayya brhaspatih | bhavato bhagavallihgavaimukhye nauciti kvacit || 74 esdm 
upapldvayitum *matim (BC : bhaktim A : satyam Ed.) esa mama kramah | 
buddher dgatam ity etad darsanam bauddham ucyate || 75 buddhah prasiddhas 
tatraikah *samkalpyeta (Codd. : samkalpeta Ed.) *suresvarah (Ed. AC : suresvara 
B) | dhydne yacchatradhartrtve likhyante kdrandny api || 76 ganapatyddayo ye 
ca saivd atyuttamdh sthitdh | tesdm murdhani likhyante devd bauddhd *aml iti 
(Codd. : amlti ca Ed.) || 77 mithyopakalpitdny evam dhydndny dlokya ddnavdh | 
sivdd utkarsavanto 'mi iti *muhyanty asamsayam (AB Ed. : muhyanti samsayam 
C). 78 evam dhydnesu siddhesu prasiddhim *lambhitesu (A Ed. : lambitesu BC) 
ca | saivatantrdnuvddena mantrdn api niyojaye || 79 uddhrtya sivasdstrebhyah 
khanddn khanddn niyojaye | mantratantrddikam krtyam yat kimcic copakalpi- 
tam || 80 bandhamoksavyavasthdydm sdstram yac ca viracyate \ tatra *tivratarah 
prajhdprakarsah (tivratarah Codd. : tivratara Ed.) *pariposakah (Codd. : pari- 
tosakah Ed.) || 81 lihgdrcanddikas tatra bandhas tdvan nigadyate | muktis tu 
sunyataiva sydd itikartavyahdrinl || 82 yajnddikd kriyd *yeyam (A : seyam Ed. 
BC) sa tatra pratihanyate | dtmd ndstlti samcintya dusyate paramesvarah || 83 
evamvidham mayd sdstram viracayya puramdara \ hrdi *pravesya (conj. : pravisya 
Codd. Ed.) bhagavadbhaktis tesdm vihanyate || 84 *sukrasydsamnidhdnam (Codd. 
: sukrasya samnidhdnam Ed.) tu tatra siddhyai *pratiksyate (Codd. : pratiksate 


The Saiva Age 

[For] he, through intuition and reason, is fully conversant with ultimate reality". 
Thus said the Atharvavedic priest [of the gods]. Then Indra respectfully asked 
him to explain the stratagem more fully. After some reflection Brhaspati said to 
Indra: "It is entirely inappropriate that it should be you that has to divert [these 
demons] from the worship of Siva's Liriga. [So I shall take on this task myself] 
My way of destroying their understanding will be this. I shall call this teaching 
Buddhist, [appropriately enough] since it will be born of [nothing more than] my 
intellect (buddhih). The well-known Buddha will be conceived therein as the sole 
lord of the gods. Even the greatest deities will be portrayed as his chowry-bearers. 
Gods that I shall call Buddhist will be depicted positioned on top of Ganapati and 
others of the highest Saiva deities. When the demons see these falsely conceived 
visualization-texts they will certainly make the mistake of thinking that these 
gods are greater than Siva. Once these texts have been established and I have 
accustomed the demons to them I shall introduce Mantras modelled on [those of] 
the Saiva Tantras (saivatantrdnuvddena) and by redacting various passages from 
these same scriptures (uddhrtya sivasdstrebhyah khanddn khanddn) I shall add 
a worthless, concocted system of [Tantric] observances involving Mantras, ritual, 
and the rest. The learned [Buddhist] literature that I shall compose to define 
bondage and liberation will be nourished by higher reasoning of an exceptional 
degree of rigour. It will explain, of course, that of these two bondage includes such 
activities as worshipping the Lihga; and liberation will be [defined as] a voidness 
[of self] that [once accepted] will subvert [their commitment to their] religious 
duties. Their sacrifices and other rituals will be opposed there; and coming to be- 
lieve [though this teaching] that there is no soul they will denigrate Siva himself 
[for teaching otherwise]. Indra, when I have composed learned teachings of this 
kind I shall insinuate them into their hearts and so put an end to their devotion 
to Siva. For the plan to succeed we have only to wait until Sukra is absent". 

Brhaspati's plan works. The demons' Saiva Guru leaves for a year to attend a 
sacrifice. Brhaspati takes on his appearance and thus disguised sets about con- 
verting them to Tantric Buddhism. They become so anti-Saiva that they can no 
longer bear even to mention the Sivalihga, let alone worship it, 502 thus making 
it possible for Siva to destroy them. 

Evidently the Buddhist Tantric scriptures that Brhaspati is represented 
here as having concocted are the Yoginitantras as typified by the Laghusamvara 
and its satellites; 503 and the fact that this understanding of the nature of the 

502 Haracaritacintdmani 13.127c-128b: Hatahprabhrti (A : taddprabhrti Ed. B) te 
daitydh sivabhaktipardhmukhdh | asahanta na lihgasya ndmdpi kim utdrcanam. 

503 That this is the Buddhism envisaged here is in keeping with another anti-Buddhist 
myth in this collection (Haracaritacintdmani, chapter 17 and SANDERSON 1995b, 
p. 94 for a summary). For there the adherents of Buddhism are said to be led by 
three demons: Heruka, Samvara (the two Vajradakas), and Adibuddha (Kalacakra). 


Genesis and Development of Tantrism 

genesis of these texts appears in a work of this kind suggests that it was com- 
mon knowledge. For the Haracaritacintamani, being concerned with the cults of 
Siva at sacred sites, is not addressed to the narrow community of the initiated 
but to the widest possible audience for a Saiva text in Sanskrit, that is to say, 
the uninitiated Saiva laity Nor can this text be seen in spite of this as reflecting 
the knowledge of a learned minority at the time of its composition. For between 
the opening and closing verses of each chapter the text is written in a rather 
unpolished style that is so similar to that of the anonymous Puranic tracts in 
praise of sacred sites that it should not be seen as a composition in the full sense 
of that term but rather as a compilation in which Jayadratha has lightly edited 
pre-existent materials of this popular genre. 504 

The redactional relation between the Yoginitantras and Saiva Tantras of the 
Vidyapitha may not, of course, have been so obvious to learned Buddhists once 
these texts had been propagated and the work of commentary undertaken, let 

17.4: mdydsambariko ndmnd herukdkhyas ca ddrunah \ adibuddhabhidhanas 
cety asurds traya asate; 17.9: vajraddkdv iti khydtau tadd herukasambarau | 
adibuddhena sahitau surdndm cakratur bhayam. Heruka here is evidently Heva- 
jra, since he is described as eight-headed, four-legged, sixteen-armed, and em- 
braced by Nairatmya (17.5). He leads the Buddhists in their war against the gods. 
He is surrounded by an army of Madhyamikas (madhyamandmdnah), followers of 
the Mantranaya (mantranaydtmakdh), bhramamohdtmakah, mithyajndndtmakdh , 
Sravakas (srdvakdtmdnah), and Buddhas copulating with their consorts (17.7-8). 
The meaning of the terms bhramamohdtmakah and mithyajndndtmakdh is not 
immediately obvious. Since it is clear from the context that they refer to dis- 
tinct groups among the Buddhists (bhramamohdtmakah kecin mithyajndndtmakdh 
pare) I take them to mean 'those who are devoted to the delusion of [the ob- 
jective existence of] non-objective cognitions') and 'those who are devoted to 
the view that [belief in this reality of] cognitions [containing the appearance 
of their objects] is false', understanding these expressions to refer to the two 
kinds of Yogacaras, those who hold mind-only with form and mind-only with- 
out form respectively to be ultimately real, that is to say Sakaravijfianavadins 
and Nirakaravijnanavadins. Classifying Mahayana Buddhists into Madhyamikas 
and these two kinds of Yogacaras and the classification of all these into those 
who follow the Mantranaya and those who do not, that is to say, those who 
follow the non-Tantric Paramitanaya, is a commonplace in the doxographical 
tradition of late Indian Buddhism; see, e.g., Advayavajra, Tattvaratndvali, pp. 
4-8; Sahajavajra, Sthitisamdsa ff. 4vl-6r2 (nirdkdrayogdcdrasthitisamdsah), ff. 
6r2-7rl (sdkdrayogdcdrasthitisamdsah), ff. 7rl-llr3 (madhyamdsthitisamdsah), 
and ff. Ilr3-18v5 (Mantranaya); Vagisvarakirti, Tattvaratndvalokavivarana, pp. 
141—142 (mantranaye ca vijndnavddamadhyamakamatayor eva pradhdnatvdt . . . ); 
Moksakaragupta, Tarkabhdsd, pp. 107-110; and Kajiyama 1998, pp. 148-151, 
504 Consider Jayadratha's own statement at the beginning of the work (1.5): 
dese srivijayesasya nivasan prerandt tayoh | caritrdni trinetrasya sdstradrstdni 
gumphaye 'While living in the land of Siva Vijayesvara I shall string together the 
deeds of the Three-Eyed [God] as I have seen them in the sacred texts, at the insti- 
gation of these two [teachers]'. 


The Saiva Age 

alone to the ordinary lay devotee of the Buddha. But the iconographical reper- 
toire, the retinue types, the style of worship and Kapalika observance, and the 
growing autonomy and diversification of the goddess, are so closely parallel to 
what we see among the Sakta Saivas that it is hard to believe that any Indian, 
learned or not, could have seen these deities and observed the practices of those 
that propitiated them without being aware of this fact. 

This must have been especially so in east India. For the Sakta tradition was 
particularly strong there, as it still is, and had deep roots in the domain of pop- 
ular religion, as is evident from such Puranas of the region as the Devlpurana, 
Brhannaradlyapurana, Brhaddharmapurana, and Kalikapurana, 505 from non- 
eastern testimony, 506 from the fact that east-Indian locations are conspicuous in 
early lists of the Sakta sacred sites, 507 and from the inscriptions and other his- 

505 See CHAKRABARTI 2001 passim. The Devlpurana (39.143-145) lists places where 
the Mother goddesses are especially present. In this list are Varendra, Radha, 
and Kamarupa: vesyasu gopabalasu tudahunakhasesu ca | plthe himavatas *calpa 
(?) *jalandhare (corr. : jalandhara Ed.) savaidise || *mahodare (?) varendre ca 
radhayam kosale pure | bhottadese sakamakhye *kiskindhe (corr. : kiskindhye) ca 
nagottame \\ malaye *kollaname (conj. : koluname Ed.) ca kancyam ca hastinapure 
I ujjayinyam ca ta vidya visesena vyavasthitah 'Those Vidyas are especially present 
among courtesans, cowherd girls, *Tudas (?), Hunas, and Khasas, in the sacred 
site of Himalaya* . . . (?), in Jalandhara, Vidisa, *Mahodara (?), Varendra, Radha, 
the capital of Kosala, Tibet, Kamarupa, the great mountain of Kiskindha, Malaya, 
*Kolla[giri] (conj.), Kanci, Hastinapura, and Ujjayinf. 

506 A verse in a Puranic passage on the calendrical festivals of Kashmir cited 
by Laksmidhara early in the twelfth century in the Niyatakalakanda of his 
Krtyakalpataru (p. 410, 11. 4-5) associates the sanguinary cult of Durga/Bhadrakali 
with the peoples of Bengal and Orissa (Ahga, Vahga, and Kalinga), the Kinnaras, 
the Barbaras, and the Sakas: evam nanamlecchaganaih pujyate sarvadasyubhih \ 
ahgavahgakalihgais ca kimnarair barbaraih sakaih 'She is worshipped in this way 
by various foreign communities, by all the Dasyus: the people of Anga, Vanga, and 
Kalinga, the Kinnaras, the Barbaras, and the Sakas'. In this list only the people 
of Ahga, Vahga, and Kalinga and the Iranian Sakas (if this reading is sound) are 
well-known. As for the Kinnaras and Barbaras, Varahamihira locates the former, 
under the synonym Asvavadana, in the east (Brhatsamhita 14.6ab: khasamagadha- 
sibiragirirnithilasarnatatodrasvavadanadanturakah), and the latter in the south- 
west (14.18c). 

507 See Sanderson 2001, p. 7, fn. 4. This is particularly clear in the case of the 
eight principal sites among the twenty-four: the eight Ksetras, namely Attahasa, 
Caritra, Kolagiri, Jayanti, Ujjayim, Prayaga, Varana/Varanasi, and Kotivarsa 
(see here p. 195), or, in a variant, Prayaga, Varana/Varanasi, Kollagiri, Attahasa, 
Jayanti, Caritra, Ekamra, and Devikota (see, e.g., citation of the Madhavakula 
in Tantralokaviveka on 29.67; Kularatnoddyota f. 13r3-4: prayaga varuna holla 
attahasa jay antika \ caritraikamrakam caiva *devikottam [corr. : devikostham Cod.] 
tathastamam). Attahasa, Kotivarsa/Devikota, Caritra, and Ekamra are all in east- 
ern India, the first two in Bengal and the last two in Orissa. The location of 
Jayanti is uncertain. It too is east-Indian if it is the Jayantipura in the Ganjam 
District of Orissa rather than that in Karnataka (Banavasi). Other east-Indian 
sites among the twenty-four are Viraja (Jajpur in Orissa), Nagara (Pataliputra, in 


Genesis and Development of Tantrism 

torical records of this period. Thus when Devapala is eulogized in an inscription 
of his son Mahendrapala it is for two achievements: his martial success and, as 
we have seen, his building of two exceptional temples, one of the Buddha and the 
other of the Saiva Goddess; Sakta Saiva deities figure strongly, as we have seen, 
in the various pious works of Nayapala detailed in the Siyan inscription: several 
Vadabhi temples for goddesses, one of them for a hill -top Carcika installed by his 
predecessor Mahendrapala, temples for the Nine Durgas, and temples for the 
[Bhairava] Hetukesvara and a Bhairava accompanied by a retinue of sixty-four 
Mothers; 508 and Madanapala, the patron of Samdhyakaranandin, is described in 
that poet's Ramacarita as having attained his success in war through the favour 
of Candi. 509 Even the Saiddhantika Prasasti from Bangarh has a Sakta context, 
its immediate purpose being to report the building by the Rajaguru Murtisiva of 
a Vadabhi temple for Carcika. 510 

[Murtisiva], being devoted to pious works, has constructed this Vadabhi temple 
which seems to embody his two halves miraculously transformed in a mountain 
of snow and a mountain of gold. I fancy that Indra's elephant, now that he can 
see the wondrous reflection of the lions [on its roof] in the waters of the heavenly 
Ganges, will recoil [in fear] and no longer drink its waters. 

That the temple is described as a Vadabhi surmounted by lions establishes that 
it is a temple of a goddess. 511 The inscription does not state explicitly that this 
goddess is a Carcika: it did not need to do so since the inscription was not doubt 
in situ. But we can infer that she was from the fact that the inscription begins 
with obeisance to her followed by two benedictory verses in her praise: 512 

Bihar), and Pundravardhana (in Bengal) among the eight Samdohas or Upaksetras 
(Nisisamcara f. 15vl [3.26]; Kubjikdmata 22.32-38), and Prsthapura (Pistapura in 
Kalinga, in the East Godavari District of Andhra Pradesh), and Rajagrha (in Bi- 
har) among the eight Upaksetras or Samdohas (Nisisamcara f. 15v3-4 [3.29]; Kub- 
jikdmata 39-46). We see the same emphasis on the east of India in the scheme of 
nine sacred sites (three Pithas, three, Upapithas, and three Samdohas) taught in 
the Nisisamcara. In the version of that text known to Abhinavagupta and his com- 
mentator Jayaratha the three Pithas are Kamarupa (Assam), Purnagiri (in the Dec- 
can), and Uddiyana (Swat). The Upapithas and Samdohas are Pundravardhana, 
Varendra, Ekamra, Devikota (all four in eastern India), Ujjayini, and Kollagiri; see 
Tantrdloka 15.83c-88. 
808 For Nayapala's foundations see here pp. 111-114. 

509 Samdhyakaranandin, Ramacarita 4.21: candicaranasarojaprasddasampanna- 
vigrahasrikam \ na khalu madanam sdngesam isam agdj jagadvijayasrih 'Did not 
the glory of world-conquest come to King Madana when, with the king of Ahga, he 
had achieved success in battle by the favour of the lotus-like feet of Candi?' 

510 SlRCAR 1983b, v. 25: teneyam himakdncandcalamahdkautuhaldvesitasviydrdhd- 
rdhavapusmativa vadabhi punydtmand nirmitd | yatsimhapratibimvam ambara- 
dhunltoyesu manye 'dbhutam drstvd samkucadahghrir adya na jaldny airdvatah 
*pdsyati (em. : pasyati Ep.). 

811 See here p. 112. 

512 om namas carcikdyai || surdsurasirahsrenipatavdsasamd jagat \ pdntu visvakrtd- 

The Saiva Age 

Obeisance to Carcika. 

May the world be protected by the dust from the feet of Carcika, worshipped by 

the creator of the universe, fragrant powder for the heads of all the gods and 


May Carcika protect the world, who at the aeon's end, garlanded with human 

skulls, with her body becoming desiccated out of anxiety at the poverty of her 

fare, thinks: "What shall I eat? If I devour this universe in a single bite, it will be 

no more than a fragment that will lodge between my teeth. What shall I drink? 

The water of [all] the seven oceans is insufficient to be visible in the hollow of my 

palm. 513 

That a Saiddhantika Guru should have built a temple for a fearsome goddess 
of this kind is compelling evidence of the strength of Saktism in the Pala realm. 
For there is nothing in the Siddhanta itself to prompt such a construction, that 
tradition generally marking itself off from the cults of such deities with their 
gruesome iconography and their ecstatic and transgressive rites. 

Indeed, as this anomalous foundation suggests, the cult of the emaciated 
Carcika seems to have been particularly well-established in the region. There 
are numerous surviving images of this goddess at or from sites in Bihar, West 
Bengal, Bangladesh, and Orissa, dating from the ninth century to the four- 
teenth; 514 she figures prominently in the east-Indian Sakta DevTpurana; 515 and 

bhyarcas carcacaranarenavah \\ damstrasamdhinilinam ekakavalam visvam tad 
asnami kim saptdmbhodhijaldni hastasusire guptdni kim piyate | ity aharadaridra- 
takulataya susyattanum bibhratl kalpdnte nrkapalamandanavidhih pdydj jagac 

513 With these verses compare those of the east-Indian poets Bhasoka and 
Umapatidhara in the anthology Saduktikarnamrta (vv. 126 and 129), compiled 
by the east-Indian Sridharadasa in 1205 under Laksmanasena. Bhasoka's being 
east-Indian is evident from his name in -oka; see the many names of this kind in 
the east-Indian anthologies Subhasitaratnakosa, and Saduktikarnamrta, Amrtoka, 
Sarigoka, Ucchoka in the inscriptions of Bengal (N.G. Majumdar 2003, pp. 179, 
27, 37, 178), and Dibboka and Rudoka in the commentary on Ramacarita 1.39. 
Umapatidhara composed the Deopara inscription of the Sena king Vijayasena (r. 
c. 1096-1159) and is reported in Merutunga's Prabandhacintamani to have been a 
minister of the Sena Laksmanasena (r. c. 1179-1206); see N.G. Majumdar 2003, 
p. 45. 

514 See Camunda (Camunda) in the Huntington Archive. For Orissa see also DONALD- 
SON 1991. 

515 See in particular DevTpurana, Patalas 7 and 9 (> Agnipurdna 135) on Camunda's 
Padamalamantra. In that Mantra Camunda is described as having her body clothed 
with an elephant hide (gajacarmapravrtasarire). This feature, which was borrowed 
from the iconography of Siva not only by Camunda but also, as we have seen, by 
Cakrasamvara and Vajravarahl, is found in most of her east-Indian images. See 
Huntington Archive, Scans 0058416 (Bangladesh), 0006042 (Itahar, North Dina- 
jpur District, West Bengal), 0013693 (findspot not recorded), 0013697 (findspot 
not recorded), 0002686 (Harsinghpur, Darbhanga, Bihar), 0000308 (West Bengal), 


Genesis and Development of Tantrism 

in early canonical treatments of the Sakta Saiva sacred sites this goddess is said 
to preside at Devlkota, 516 Pundravardhana, 517 both in Varendri, and Ekamra 
(Bhubaneswar) in Orissa. 518 In the first she has the name Karnamoti, 519 accord- 
ing to the Nisisamcara, Picumata, and Kubjikamata, and Bahumamsa according 
to the Skandapurana-Ambikakanda. 520 In the other two she is called Camunda. 

0013061 (Dighapatiya, Natore District, Bangladesh), 0002607 (Munger [Monghyr], 
Bihar), 0013063 (Bangladesh), 0013062 (Mahatore, Dinajpur District), and 0013476 
(Vikramapura, Dacca District, Bangladesh); also AIISPL Ace. no. 32782 (Advahati, 
Burdwan, West Bengal). It is not generally seen in images of Camunda from other 
regions. An exception is a fine sculpture at Khajuraho (AIISPL Ace. no. 45199) 
from the Chandella period (c. 900-1150). It is perhaps to be introduced by emen- 
dation into the description of Camunda's icon in Agnipurdna 50.21c-23b: camunda 
kotaraksT sydn nirmamsa tu trilocand || nirmamsa asthisdrd va urdhvakesl krsodarl 
| *dvipacarmadhard (dvipa conj. : dvipi Ed.) vdme kapalam pattisam kare || sulam 
kartri daksine 'syah savdrudhdsthibhusand. 

516 See here p. 112. 

517 Nisisamcara f. 18v2-3 (4.35-36): camundeti ca *vikhydtd (em. : vikhyd Cod.) devya 
va *pundravardhane (corr. : punda Cod.) | mahdbaldkulotpannd khatvdhga- 
karasobhitd || 36 bhuktimuktikard devya samdohaksetrasamsthitd | kumbhdkhyo 
ksetrapdlas ca tasmin ksetre vyavasthitah; Kdlikulakramdrcana f. 21vl: HRIM 
srIm sri*pundravardhanamahopaksetre camunda-ambapada {pundra corr. 
: punda Cod.). 

518 Nisisamcara f. 31rl-2: *ekamre (em. : ekatye Cod.) *samsthito (corr. : samsthita 
Cod.) devi kirti*vaseti (corr. : taseti Cod.) *kirtitah (corr. : kirtita Cod.) | ca- 
mundaya (corr. : camundaya Cod.) samayu*ktah (corr. : ktam Cod.) sthana- 
balisamanvi*tam (corr. : tah Cod.); Kubjikamata 15.28—30: vartamdnikakalpe tu 
ekamrakavanantagah \ kapalisa*kulesanacarnundacakramadhyagah (kulesana 
corr. : kulesanam Ed.) || 29 srikulesvaradevasya hrtpadme 'stadale sthitah | 
isanakramayogena srstimdrgavalambikah || 30 karnikayam sthito devas catuska- 
parivaritah | raktdkardldcandaksimahocchusmasamanvitah; Kularatnoddyota f. 
16r2 (3.140c— 142b): ekamrakavandntasthd utpannd<h> paramesvari || 141 
kapdlisasamopetds camunda*cakramadhyagdh (corr. : scakra Cod.) | pithasthdnd- 
srayodbhutds catasro 'nyd<h> pardmbike | 142 raktd kardld canddksi ucchusmeti 

519 Karnamoti is listed as a synonym of Camunda in Amarakosa 1.1.92 (see here 
p. 231). The name appears for Camunda in the series of eight Mother goddesses 
when these are given as the deities of the seven sets of sounds of the Sanskrit syl- 
labary plus KSA in Siddhayogesvarimata 16.41c— 43c: kavarge samsthita brdhmi 
cavarge caiva vaisnavT || mdhesvari tavargasthd ydmyd pujyd ta-m-ddind | kaumdri 
sarpavalayd pddyenaitdm prapujayet || yavarge vdsavT tatra karnamoti sa-m-ddind 
| krodhe *jheyd (conj. : seyd Ed.) para saktir aghoresT 'Brahml is in the gutturals, 
Vaisnavi in the palatals, Mahesvari in the retroflexes, and Yamya in the dentals. 
He should worship snake-bangled Kaumari with the labials. Aindri is in the semi- 
vowels and Karnamoti (= Camunda) in the sibilants. Know that the goddess in ksa 
is the supreme Power Aghoresvarf. The origin of the name is unknown, the common 
interpretation 'Ear-pearl' being implausible since it fails to account for the retroflex 


520 Skandapurdna-Ambikdkhanda 171.109, 112, 124 This name is probably an epithet 
that served as this Karnamotl's personal name and so does not indicate a different 
goddess. The epithet, meaning 'having much meat', no doubt refers to her insa- 


The Saiva Age 

Of these sites Devlkota appears to have been of special importance from 
early times. The Mddhavakula refers to it simply as Sripitha, that is to say, as 
the Seat [of the Goddess]; 521 and the Skandapurana-Ambikakhanda describes 
it 522 as a city originally fashioned by Brahma where this goddess and the other 
Mothers who accompany her were created by Siva and the other gods from 
their own bodies in order to destroy the demons who had seized it. After the 
city has been freed Siva declares that henceforth it will be the Mothers' sacred 
abode, 523 that he will reside here with them as Hetukesvara, 524 and that they 
will be worshipped following ritual procedures taught in Tantras that will be 
composed for this purpose by the grateful gods. The titles of these Tantras of 
the Mothers {mdtrtantrani), which are listed in the narrative, reveal them to be 
Yamalatantras, headed by the Brahmayamala. 525 

tiable appetite for animal sacrifices. The alternative, that it means 'fleshy', that 
is to say, full-bodied, is highly implausible, since she is described here as the de- 
stroyer of the universe and as having a hideous form (171.108c-109: tato devo 'srjad 
devim rudranlm mataram subham | vikrtam rupam asthaya dvitlyam api mataram 
| namna tu bahumdmsdm tarn jagatsamhararupinim 'Then the deity [Siva] em- 
anated the fine Mother goddess Rudram, and, taking on a hideous form, a second 
Mother, the [well-known goddess] called Bahumamsa, who embodies the destruc- 
tion of the universe'. 
621 See here p. 192 and Tantraloka 29.60cd. 

522 Skandapurana-Ambikakhanda 171.78-137, referring to Devlkota under its name 
Kotivarsa. See here p. 113. 

523 Skandapurana-Ambikakhanda 171.120c-121b [Siva addresses the Mothers]: bha- 
vatlnam idam sthanam kotlvarsam iti srutam | bhavisyati jagatkhyatam sar- 
vapdpapramocanam 'This place known as Kotivarsa will be yours, famed through- 
out the world, with the power to free from any sin'; 171.133cd: kotlvarsam idam 
sthanam mdtrnam priyam uttamam. 

524 SkandapuranaAmbikakhanda 171.121c-122b [Siva addresses the Mothers]: aham 
hetur hi yusmakam yasmdt srsta mayaiva ca || herukesvaranamnaham sthasydmy 
atra varapradah \ yusmabhih saha vatsyami nayakatve vyavasthitah \\ yas tu 
yusman maya sardham vidhivat pujayisyati | sarvapapavimuktatma sa param 
gatim dpsyati 'Because I am your cause (hetuh) and it was I that created [you], I 
shall be present here to bestow boons with the name Hetukesvara. I shall dwell 
here with you as your leader. Whoever correctly worships you with me will be freed 
from all sins and attain the highest goal'. 

525 Skandapurana-Ambikakhanda 171.127-132b [Siva addresses the Mothers]: aham 
brahma ca visnus ca rsayas ca tapodhanah | mdtrtantrani divyani matryajnavidhim 
*prati (conj. .param Cod.) || 128 punyani prakarisyamo yajanam yair avapsyatha \ 
brahmam svayambhuvam caiva kaumaram yamalam tatha || 129 sarasvatam ca 
gandharam aisanam nandiyamalam \ tantrany etani yusmakam tathanyany sa- 
hasrasah || 130 bhavisyanti nara yais tu yusman yaksyanti bhaktitah \ naranam 
yajamdnanam varan yuyam pradasyatha || 131 divyasiddhiprada devyo di- 
vyayoga bhavisyatha \ yds ca naryah sada yusman yaksyante sarahasyatah || 132 
yogesvaryo bhavisyanti rama divyaparakramah 'I, Brahma, Visnu, and the as- 
cetic sages will compose excellent and holy Matrtantras for the rites of the wor- 
ship of the Mothers, by means of which you shall receive offerings. The Brah- 


Genesis and Development of Tantrism 

Moreover, it is probable that some at least of the surviving east-Indian im- 
ages of the emaciated goddess reproduce the iconography of this important local 
form. An image of Carcika from the Dinajpur District of Bangladesh, in which 
Devikota was located, shows the goddess seated beneath a banyan tree; 526 and 
we see the same in an image from an unrecorded site in West Bengal. 527 In both 
images severed human heads are attached by their hair to the tree's branches, in- 
dicating that the site of this tree is a cremation ground, since cremation grounds 
were also places of execution. 528 Now, in the tradition of the Picumata and 
the Nisisamcara each of the major Sakta sites is a cremation ground with its 
own distinctive sacred tree; and in the case of Kotivarsa/Devikota this is in- 

maydmala, the Svayambhuydmala, the Skandaydmala, the Sdrasvataydmala, the 
Gdndhdraydmala, the Isdnaydmala, and the Nandiydmala: you shall have these 
Tantras and others in thousands, and with them men will sacrifice to you in devo- 
tion. You will grant boons to men who sacrifice to you. Being goddesses of celestial 
power you will bestow celestial Siddhis. And women who sacrifice to you regularly 
with the secret [ritesl will become Yogesvaris, women of celestial might'. On the 
list of Yamalatantras in this passage and its relation to lists of such texts in the 
Vidyapitha see SANDERSON 2001, pp. 6-7, fn. 4. The Brahmaydmala, also called 
Picumata, teaches the worship of Bhairava as Hetuka surrounded with the God- 
dess by eight Viras and twenty-four Yoginis in its eightieth chapter (f. 306r2-3; 
80.32-33): hetukam devadevesam kapdlakrtabhusanam | vlrdstakayutam madhye 
devadevam parodayam || kdldgnivdyusamyuktam adhordhvakrtasamgatim | nyaset 
svarupabhdsvantam tato yogiganam nyaset. It is striking that this reference to 
Hetuka, presumably the Bhairava of Devikota, is found in a chapter which is dis- 
tinguished by being one of the very few passages in the Vidyapitha that departs 
from the Tantric norm by containing material of the Puranic type, the subject 
which gives it its title being a myth of the origin of the skull-bowl and skull-staff 

526 Pala period; black stone; 9 inches in height; now in the Varendra Museum in Ra- 
jshahi: Huntington Archive , Scan 0013117. 

527 Sena period; black stone; 25.75 inches in height; now in the National Museum, New 
Delhi: Huntington Archive Scan 0000308. 

528 See, e.g., Kumdrasambhava 5.73cd; Kathdsaritsdgara 18.130d; Rdjatarahginl 
2.79-84; Picumata 3.32d-93, describing the depiction of the cremation ground 
at Prabhasa: tato nimbam samdlikhet | saptaddlam mahdbhimam citibhih 
prajvalantibhih \ ekaikasmim likhet dale nagnam udbaddhakam naram 'Then he 
should depict a Nimba tree with seven branches, most frightening with the burning 
pyres [around it]. On each branch he should draw a naked hanged man'; 15.16: 
krsndstamydm caturdasydm savam grhya tha sddhakah | udbaddham siilaprotam 
vd aksatdhgam tu ddrakam; Jayadrathaydmala, Satka 3, Yoginisamcdraprakarana 
8.71c-72b, describing the depiction of cremation grounds: ydmyddyair nairrtdntais 
tu disair vrksdn samdlikhet || udbaddhanarapracchanndn; Vajragarbha on Hevajra 
1.7.21 (dhvajam sastrahatam caiva) quoted in SNELLGROVE 1959, Pt. 1, p. 71, n.: 
rgyal mtshan ni rgyal pos rkun po la sogs pa skyes pa 'am bud med ga' zhig chad 
pas bead de lus mtshon gyis dral nas ro shing la dpyangs pa'o 'a dhvajah is a corpse 
of some man or woman guilty of theft or some other crime whom the king has had 
executed with the sword, which has then been hung up on a tree [in the cremation 


The Saiva Age 

deed the banyan (vatavrksah). 529 This strongly suggests that the local Carcika 
of Devikota may have been multiplied in the manner of the Nataraja of the Tamil 
country, which though originally the deity of Cidambaram was established in sec- 
ondary forms in temples throughout the region. We may note also that most of 
the surviving east-Indian Carcikas hold the trident, often as the most conspic- 
uous of their held attributes. Both the Picumata and the Nisisamcara specify 
this as the weapon distinctive of the Karnamoti of Devikota, and the Skanda- 
purana-Ambikakhanda says that it is because the goddess of this place slew 
the demons with her trident here that the site contains a sacred bathing-place 
called Sulakunda 'the pond of the trident' and that anyone who drinks its water 
{sulodakam) after doing obeisance to her will be safe from all harmful beings 
(171.124-125). The Picumata too refers to this Kunda. 530 

Finally, the pre-eminence of the emaciated goddess in the Saktism of eastern 
India during this period is strongly underlined by the fact it is she that the Bud- 
dhists of the cult of Cakrasamvara chose to represent supine beneath the right 
foot of Samvara and Vajravarahi as the female representative of the Sakta Saiva 

In textual references to that Buddhist icon she is generally called Kalaratri. 
But there can be no doubt about her identity. For (1) she is called Carcika 
in the Vajravarahisadhana of the Siddha Luyi, 531 and Camunda in a Kalpa of 
the Abhidhanottara and in the anonymous Trayodasatmakavajradakinivajra- 
varahisadhana, which is based upon it; 532 (2) Carcika is called Kalaratri in a 

529 See here p. 112. That the sacred sites are the cremation grounds {smasanam) 
of the places listed is clear from the context in the Picumata, that (3.8-127) 
being a description of the nine cremation grounds that must be installed in 
the initiation Mandala (mahamandalam), one at the centre (Prayaga) and eight 
around the periphery (Varanasi, Viraja [Jajpur in Orissa], Kollagiri [Kolhapur 
in Karnataka], Prabhasa [in Kathiawar], Ujjayim [in Malwa], Bhutesvara [in 
Mathura?], Ekamraka [Bhubaneswar in Orissa], and Kotivarsa). It is also clear 
from the account of Kotivarsa given in the SkandapuranaAmbikakhanda, since 
that prophesies that the site will become a great cremation ground (171.133c-134b): 
kotivarsam idam sthanam matrnam priyam uttamam || smasanam pravaram 
divyam bhavisyati sukhapradam. 

530 Picumata f. 8r3 (3.119c-121b): isane tu disdbhdge kotivarsam. prakalpayet \\ 120 
vatam tatra samalikhya tatra sulodakam likhet | diksu caiva vidiksu ca sulaprota 
likhet tatha || 121 sula tasydgrato likhya kundasyaiva mahatape. It appears from 
this that the pond (kundam) was also known as the Sulodaka. 

531 Guhyasamayasadhanamala f. llrl-2: vamabahustanamandalahrdayasambhava- 
*militadaksinahghrim (em. : milita \ daksindmghri Cod.) carcika<m> rakta<m> 

532 Abhidhanottara, Patala 56, A f. 173v2: padatalakrantabhairavacamunda 'treading 
on Bhairava and Camunda with the soles of her feet'; Trayodasatmakavajradakinl- 
vajravarahisadhana in Guhyasamayasadhanamala, f. 78r4-5: padakranta*krta- 
sambhucamundam (em. : krtam | sambhuscamundam Cod.). For the full visualiza- 


Genesis and Development of Tantrism 

verse by the east-Indian poet Bhasoka; 533 and (3) the goddess beneath the foot of 
Samvara/Vajravarahi is depicted as emaciated, with sunken eyes and withered 
breasts, holding a skull-bowl and chopping knife in her two hands. 534 The ema- 
ciated Carcikas of our surviving images have four, six, eight, or ten arms, but the 
skull-bowl (kapdlam) and chopping knife (kartrikd) are indeed among their four 
primary attributes, the other two being the trident and a severed head. 535 The 
goddess beneath the right foot is, as it were, the east-Indian Carcika reduced to 
essentials: the emaciated body, the red colour, and only two arms, brandishing 
what were felt to be her two most basic attributes. 

It is inconceivable, therefore, that east-Indians, for whom Sakta Saivism was 
so central, then as now, would not have been conscious of the Sakta Saiva guise 
of this new Buddhism; and it is equally inconceivable that they would have been 
blind to the fact that the humilated goddess supine beneath Samvara's and Va- 
jravarahi's feet was the pre-eminent goddess of the east-Indian Sakta tradition. 
Clearly the east-Indian Buddhists who developed this iconography chose this 
goddess precisely because she occupied so prominent a position in that tradition 
and therefore would be instantly recognized. 

In explanation of why this profound transformation of Buddhism occurred, 
we might be tempted to say that Buddhism was simply yielding ever more com- 
pletely to the Sakta Saiva religious tradition then dominant in the region, failing, 
as it were, to maintain its original purity in the face of this external pressure and 
the concomitant expectations of its patrons. This was perhaps how the matter 
would have been represented by the Sravakayanists; and no doubt there is some 
truth in this assessment, since it is extremely unlikely that east-Indian Bud- 
dhists would have chosen to develop this new manifestation of their religion if 
Sakta Saivism had not become the pre-eminent religious idiom of the region. But 

tion text of which this is part see ENGLISH 2002, p. 407, n. 207. 

533 Saduktikarnamrta 126. For the east-Indian character of names in -oka see here 
p. 227. 

534 For this depiction see two stone sculptures from Ratnagiri in Orissa (LlNROTHE 
1999, figs. 198-202), two bronzes, one from Vikramasila and the other from an 
unrecorded site in eastern India (Linrothe 1999, figs. 206-208), a Kashmirian 
bronze (Pal 1975, Plate 64a,b; Linrothe 1999, fig. 211; Huntington Archive Scan 
0059531), some early Tibetan bronzes (Linrothe 1999, figs. 213-214), a Nepalese 
bronze of the fourteenth century (Pal 2003, fig. 31), a Nepalese bronze dated 1772 
(Reedy 1997, fig. N299), a painting from Khara-khoto, before 1227 (Rhie and 
Thurman 1991, fig. 92), and a Nepalese painting of the early seventeenth century 
(Kreijger 1999, p. 53). In some Tibetan paintings Kalaratri's emaciation is absent 
(e.g., Pal 2003, fig. 117; Kossak and Singer 1998, fig. 43; Rhie and Thurman 
1991, fig. 69.2); but that this is a secondary development can be inferred from its 
much more restricted occurrence. 

535 See Camunda (Camunda) in Huntington Archive. 


The Saiva Age 

the iconography of the humiliation of Carcika and Bhairava and the extensive 
learned literature that developed around the kernel of the Yoginitantras alert us 
to the fact that those who created and refined this tradition saw the matter in an 
entirely different light. In their view they were not succumbing passively to an 
alien influence. Fully conscious that they were assimilating the dominant Sakta 
Saiva idiom of the region, they justified their doing so as a means of converting 
non-Buddhists, taking their practices and encoding them with Buddhist mean- 
ing so that outsiders could rise effortlessly through what was familiar to them to 
what would save them, a view exactly reflected in Jayadratha's myth of the com- 
pilation of anti-Saiva iconography, Sakta Saiva liturgy, Mantras, and Buddhist 
doctrine as a means of luring devout Saivas away from their faith. 

For while the learned literature of Tantric Buddhism claims with sincere 
conviction that its special methods are designed for exceptionally able aspirants 
within the Buddhist fold, 536 its point of entry, namely initiatory introduction be- 
fore the Mandala, was designed to facilitate the recruitment of those outside it 
and to this end access was rendered as easy as possible. Thus in the seventh 
century the Mahavairocandbhisambodhi sets out a number of qualities to be 
sought in candidates but states that if even only one of these is present there is 
no need to investigate further; 537 and in the eight century the Sarvatathagata- 

536 See, for example, the doctrine of the four points of superiority of the Tantric form 
of the Mahayana, the Mantranaya, over the non-Tantric Way of the Perfections 
(paramitanayah) asserted in the *Nayatrayapradipa by an author whose name ap- 
pears in the Tenjur as Tripitakamala, an implausible name, perhaps an error for 
Tripitakamalla (Tshul gsum gyi sgron ma, f. 16v3: de yang pha rol tu phyin pa'i 
theg pa chen po dang don gcig pa las de'i khyad par gang dag yod pa de brjod par 
bya'o | don gcig nyid 'ang ma rmongs dang | thabs mang dka' ba med phyir dang | 
dbang po rnon po'i dbang byas pas | sngags kyi bstan bcos khyad par phags 'More- 
over, although there may be no difference in the goal [of the Mantramahayana] 
from that of the Paramitamahayana the points that distinguish [the former] should 
be statedf. This has been done done in the following verse]: "Though the goal is 
one and the same the Mantrasastra is superior (1) because it is free of delusion 
[on the path], (2) because it offers many methods [for reaching the goal], (3) be- 
cause it is free of difficulties, and (4) because only those with the highest capacity 
are qualified [to undertake it]'". The Sanskrit of the verse is preserved through 
citation (without attribution) in the Tattvaratnavali of Advayavajra (p. 8) (A), the 
Sthitisamasa of his disciple Sahajavajra (f. Ilv2 [6.5]) (B), and the anonymous 
Subhasitasamgraha (part 2, p. 31) (C): ekarthatve 'py asammohad *bahupayad 
(AB Tib. [thabs mang] : vajropayad C) aduskarat | tiksnendriyadhikarac ca 
mantrasastram visisyate. It has also been cited by Ka ro pa (Karopa?), wrongly 
attributing it to a *Pradlpoddyotanatantra (sgron ma gsal ba'i rgyud), in his com- 
mentary on the Caturmudranvaya (Mathes 2008, p. 96). According to the view of 
some, as reported by Gzhon nu dpal, Ka ro pa was another disciple of Advayavajra 
(Blue Annals, pp. 842-843, 847-849, reported by Mathes [2008, p. 89] as saying 
that he was a disciple of Advayavajra's disciple Vajrapani). 

537 rNam par snang mdzad chen po mngon par byang chub pa'i rgyud (Mahavairocana- 


Genesis and Development of Tantrism 

tattvasamgraha goes so far as to prohibit the application of any criteria for dis- 
tinguishing between those who are and are not worthy. Furthermore, it makes 
this open-door policy absolutely clear by specifying those to whom introduction 
before its Mandala is intended to appeal: 538 

Next is [the topic of] the detailed procedure that begins with the entry of Va- 
jra disciples into this Great Mandala of the Vajradhatu. In this the first step is 
entry in as much it is the means of rescuing all persons without exception and 
of bringing about the accomplishment of the highest joy for the benefit of all. 
With regard to this entry before the Great Mandala [the officiant] need not ex- 
amine candidates to determine who is and is not worthy. Why is that? Venerable 
Tathagatas, there are (1) people who have commited great sins. By seeing and 
entering this Great Mandala of the Vajradhatu they will be freed of all the bad re- 
births [that would be the consequences of those sins]. 539 Venerable [Tathagatas], 

bhisambodhitantra), f. 162v4-6: de nas de yi phyi de nyin | slob ma dad cing rigs 
btsun pa | de bzhin dkon mchog gsum la dad | zab mo yi ni bio dang Idan | spro ba 
che zhing tshul khrims Idan | bzod dang Idan zhing ser sna med | dpa' la yi dam 
brtan pa ni \ bcu 'am brgyad dam bdun nam Inga \ gcig gnyis bzhi las lhag kyang 
rung | dpyad mi dgos par gzung bar bya 'Then, the next day, he should assemble 
candidates (1) with faith, (2) of good family, (3) with belief in the Three Jewels, (4) 
with deep understanding, (5) with great energy, (6) adhering to moral conduct, (7) 
patient, (8) free of envy, (9) intrepid, and (10) steadfast in their observances. They 
are acceptable without need for [further] examination if they have [all] ten, or eight, 
seven, five, one, two, four, or more [of these qualities].' 

538 Sarvatathdgatasamgraha, sections 210-213: athatra vajradhdtumahdmandale 
vajrasisyapravesddividhivistaro bhavati | tatra prathamam tdvat praveso bhavaty 
tra mahdmandalapravese pdtrdpdtrapariksd na kdryd | tat kasmdd dhetoh | 
santi bhagavantas tathdgatdh kecit sattvd mahdpdpakdrinah | ta idam vajra- 
dhdtumahdmandalam drstvd pravistvd ca sarvdpdyavigatd bhavisyanti | santi ca 
bhagavantah sattvah sarvdrthabhojanapdnakdmagunagrddhdh samayadvistdh 
purascaranddisv asaktdh | tesdm apy atra yathdkdmakaraniyatayd pravistdndm sa- 
rvdsdparipurir bhavisyati \ santi ca bhagavantah sattvah nrttagdyahdsyaldsydhd- 
ravihdrapriyatayd sarvatathdgatamahdydndbhisamayadharmatdnavabodhatvdd 
anyadevakulamandaldni pravisanti | sarvdsdparipurisamgrahabhutesu niruttara- 
ratipritiharsasambhavakaresu sarvatathdgatakulamandalesu siksdpadabhayabhi- 
td na pravisanti | tesdm apdyamandalapravesapathdvasthitamukhdndm ayam eva 
vajradhdtumahdmandalapraveso yujyate sarvaratiprityuttamasiddhisukhasau- 
manasydnubhavandrtham sarvdpdyapratipravesdbhimukhapathavinivartandya 
ca | santi ca punar bhagavanto dhdrmikdh sattvah sarvatathdgatasilasamddhi- 
prajhottamasiddhyupdyair buddhabodhim prdrthayanto dhydnavimoksddibhir 
bhumibhir yatantah klisyante | tesdm atraiva vajradhdtumahdmandalapravesa- 
mdtrenaiva sarvatathdgatatvam api na durlabham kim ahga punar anyd siddhir 

539 rp^g d oc t rme that the mere sight of the Mandala destroys all one's sins is seen here 
in section 900: tato yathdvan mukhabandham muktvd mahdmandalam darsayet 
| mandale drstamdtre tu sarvapdpair vimucyate 'Then after duly removing the 
blindfold he should show him the Great Mandala. As soon as he has seen it he 
is freed of all his sins'. But it is much older. It is already found in the Mahd- 


The Saiva Age 

there are (2) people who are attached to every [kind of] wealth, food, drink, and 
other sense objects, who are [therefore] averse to [submitting to] the rules [of the 
initiated] (samaydh) and incapable of such disciplines as the Preliminary Obser- 
vance (purascaranam). 540 If they enter this [Mandala] they too will have all their 

manivipulavimdnasupratisthitaguhyaparamarahasyakalpadhdranl, which may be 
the earliest Buddhist text teaching consecration in the context of introduction to 
a Mandala, here with the peculiarity that consecration precedes entry, while in 
the later tradition entry precedes consecration: f. 53vl-5 (Tib. f 384v7): tatah 
anena mantrendbhisincya pravesayet: OH MANIVIPULASUPRATISTHITA*SIDDHE 


: SAMBHARA Cod.) *HUM HUM (Cod. : HUM Tib.) | yathdbhisiktamdtras 
ca sarvapdpdvarandni purvajanmasamjdtdni karmdvarandni visuddhdni bha- 
vanti sarva*suddhiparigrhito (suddhi em. : suddha Cod.) bhavati sarvatathd- 
gatddhisthitah sarvatathdgatdbhisiktah 'Then he should introduce him into the 
Mandala after consecrating him with the Mantra om manivipulasupratisthita- 


SAMBHARA HUM HUM. Merely through this consecration the obscurations of all his 
sins, the obscurations of his actions committed in previous lives, are eliminated. He 
possesses all purity. He has been entered-and-empowered by all the Tathagatas. 
All the Tathagatas have consecrated him'. According to the Zhen Yuan Catalogue 
of a.d. 800 (T 2157-935a:26) the Chinese translation of this text (T 1007) was pre- 
pared by an unknown translator of the Liang dynasty (503-557). However, I do not 
yet know if this passage is found in that translation. 
540 This ig the practice otherwise known as purvasevd. It consists of a high number 
of repetitions of a Mantra along with ascetic restraints by means of which the 
practitioner qualifies himself to undertake procedures that require its use. See, 
e.g., Mahjusriyamulakalpa, p. 236: ddau tdvat parvatdgram druhya vimsallaksd- 
ni japet | purvasevd krtd bhavati | kslrdhdrena maunind ndnyatra mantragata- 
cittena trisaranaparigrhltena utpdditabodhicittena ca posadhasilasamvarasamd- 
ddpandbodhisattvasamvaraparigrhitena japtavyam \ tatah karmdni bhavanti 'Be- 
fore [beginning the Kalpa] he must first climb to a mountain top and [there] re- 
peat the Mantra two million times. [Thus] the Preliminary Service [of the Mantra] 
will have been accomplished. He must repeat the Mantra while sustaining himself 
with [nothing but] milk, maintaining silence, with his mind fixed on the Mantra 
and nothing else, after taking the three Refuges, having formally resolved to attain 
the Awakening, and having taken up the Posadha fast, the restraint of morality, 
and the restraint of a Bodhisattva. [Only] then can the rituals be undertaken.' 
This, barring the specifically Buddhist vows, is exactly as prescribed in the Saiva 
Mantramarga, where, as here, the terms purvasevd and purascaranam / 'pur ascary a 
are standard and synonymous. See, e.g., Nisvdsaguhya, f. 80v3: japamdna-m eva 
mdsena purvasevd krtd bhavati 'By repeating the Mantra for a month the Pre- 
liminary Service will have been accomplished'; and Ksemaraja Svacchandoddy- 
ota ad 7.104cd: purascaryd prathamam eva mantragrahapurvam vratam niyata- 
japddikaranam 'The purascaryd is the observance that follows immediately after 
receiving the Mantra. It is to do a fixed number of repetitions [of that Mantra] with 
certain other [requirements].' Living on a diet of milk and maintaining silence is 
also a standard feature of Saiva Mantra observances; see, e.g., Nisvdsaguhya f. 81r4: 
dasdham kslrdhdrena japtavyah kdlamrtyum jayati; f. 82vr4: naktdsi ksirdhdro 
vd maunena tu japed yas tu | sa sivo 'bdena mdnavah; f. 84v6: anena mantrena 
ksirdhdro samvatsaram japet. 


Genesis and Development of Tantrism 

hopes fulfilled in accordance with their desires. Venerable [Tathagatas], there are 
(3) people who cannot grasp the nature of the understanding of the Mahayana of 
all the Tathagatas because they are attached to dancing, singing, joking, amuse- 
ments, and the pleasures of eating, and [so] take initiation before the Mandalas of 
otherf, non-Buddhist] families of deities. Being afraid of the moral regulations [of 
Buddhism] they do not enter the Mandalas of the family of all the Buddhas, which 
comprise the fulfilment of all aspirations, which bestow the highest happiness, de- 
light, and joy. It is for these too, who are inclined to enter the way of Mandalas 
that lead to bad rebirths, that this entry into the Mandala of Vajradhatu is ap- 
propriate, so that they may experience every happiness and delight, the highest 
Siddhi, joy, and contentment and be turned aside from the path that leads them 
to enter all [Mandalas that result in] bad rebirths. Venerable [Tathagatas], there 
are also (4) pious persons, who seek the Buddhas' enlightenment by means of 
the morality (silam), concentrations (samadhih), and wisdom (prajna) of all the 
Tathagatas but who experience hardship as they strive to attain the levels of the 
meditations (dhyanam), liberations (vimoksah), and the other [states on the path 
taught in the Paramitanaya]. They will easily attain All-Buddha-hood without 
difficulty in this very life (atraiva), all the more so other Siddhis, simply by enter- 
ing this Mandala of Vajradhatu. 

Thus the text offered Mandala initiation not only to Buddhists, and in par- 
ticular to those who had found themselves unable to progress on the exacting 
path of the Paramitanaya, but also to sinners and sensualists regardless of their 
religion, and, most important in the present context, to outsiders who had al- 
ready taken a non-Buddhist Tantric initiation or might otherwise be expected do 

The Sarvatathagatatattvasamgraha does not tell us whether it has particu- 
lar kinds of non-Buddhist Tantrics in mind. We can only guess from the character 
of the initiation ceremony, with its emphasis on possession, and the cult to which 
initiation leads, with its erotic and sensual elements, that Sakta Saivas must 
have been intended. Later sources, however, do make clear that it is indeed the 
non-Buddhist followers of the kinds of practice being adapted by the Buddhists 
that are in mind. Thus Anandagarbha, the period of whose activity, though not 
yet narrowly determined, may be assigned to the ninth century, 541 attempting 

541 The dating of Anandagarbha in the ninth century seems probable solely on the 
grounds of the range of his exegesis, which covers the Yogatantra systems of 
the Sarvatathagataattvasamgraha (his Sarvavajrodaya, his commentaries on the 
Sarvatathagatatattvasamgraha [Toh. 2511]), the Paramadya (his commentary 
[Toh. 2512]), the Mayajala (his commentary [Toh. 2513]), Guhyasamaja (his com- 
mentary [Toh. 1917]), and the Sarvabuddhasamayogadakinijalasamvara (his com- 
mentary on the Sarvakalpasamuccaya [Toh. 1662]). In the last of these Tantric 
systems we also have in Sanskrit but not in Tibetan translation his Vajrajvalodaya 
nama sriherukasadhanopayika in a codex photographed by Rahula Sarikrtyayana 


The Saiva Age 

in his commentary on the Guhyasamdjatantra to explain the extraordinary fact 
that the place where the Buddha is said to have been residing at the time that 
he revealed this Tantra is the vaginas of the goddesses, declares: 542 

If it is asked why he was residing in their private parts, the answer is [that this 
is] in order to bring it about that those devoted to the Tantras of Visnu and the 
other [gods], who have not yet abandoned [their attachment to] the objects of the 
senses, may come through desire itself to delight in the abandoning of desire. For 
they seek to attain the Siddhis of such [gods] as Visnu by resorting to women, 
and using such [offerings] as beef and urine. Those engaged in the quest for the 
Siddhis taught by these [gods do indeed] copulate with women [for this purpose]. 
For [it is said in their texts]: "Visnu is Bhagavan ['the possessor oibhaga-'] in that 
he resides in the genitals (bhaga-) of women. He is called Narayana [for the same 
reason,] because [by residing there] he gives pleasure to men". 543 

in the Ngor monastery in Tibet which comprises apart from this work forty-one 
items pertaining to the cult of Hevajra (Isaacson 1999). The dating is supported 
by the tradition (Blue Annals, p. 373) that he was a pupil of Dipankarabhadra, who 
was a pupil of Buddhajnana, a contemporary of king Dharmapala (r. c. 775-812) (see 
here p. 93). 

542 gSang ba 'dus pa'i dka' grel, f. 4r3-5: ci'i phyir de dag gis gsang ba la bzhugs she 
na | smras pa khyab 'jug la sogs pa'i rgyud la mngon par dga' zhing yul yongs su mi 
spong ba mams ni 'dod chags kyis 'dod chags spong ba 'di la dga' ba bskyed par bya'i 
phyir te \ 'di Itar bud med bsten pa dang *ba sha dang (conj. :bshad Derge, Cone, 
Ganden) gci la sogs pa bsten pas khyad jug la sogs pa bsgrub par 'dod cing | des 
bstanpa'i dngos grub tshol pa la zhugspa de dag btsun mo'i gsangpa la mngon par 
jug par 'gyur te \ de yang \ bha ga legs Idan khyab jug ste \ bud med kyi ni mdoms 
na gnas | mi mams dga' bar byed pas na | des na sred med bu zhes bya zhes bshad 

543 The unknown author of this verse intends a nirvacanam of narayanah. A nirva- 
canam is a kind of semantic analysis that explains why a word is appropriate to 
that to which it is applied (anvartha-). When this is not thought to be adequately 
revealed through ordinary grammatical analysis one may resort to an analysis in 
which the meaning sought is discovered by deriving one or more of a word's syl- 
lables from a verbal root that resembles it in sound. See the analysis of Yaska's 
statement of this principle in Kahrs 1998, pp. 35-39. In this case the name is 
made to mean 'he who gives pleasure to men'. The first component in this analy- 
sis of narayanah was evidently nara-, understood as either as 'sons of man' (nara-) 
by Astadhyayl 4.3.120 (tasyedam; cf Manusmrti l.lOab in another nirvacanam of 
narayanah: apo nara iti prokta apo vai narasunavah), or as 'men' {nara-) by ap- 
plication of Astadhyayl 6.3.136 (anyesam api drsyate) to account for non-standard 
lengthening of the first vowel. For these two alternatives see Kulluka on Manusmrti 
l.lOab and Medhatithi on the same for the second. Since ay a- can mean 'good for- 
tune', I speculate that the author found his meaning by deriving the last syllable, 
-na, from y>if- 'to lead [to]', arriving by this artifice at 'he who leads men to good 
fortune, i.e. happiness' (naran ayam sukham nayatlti narayanah), the substitu- 
tion of n for n being caused by the preceding r. The artificial derivation of -na 
from yjnl- is seen in the semantic analysis of samanah for the fourth of the five vi- 
tal energies implicit in, e.g., Nisvasanaya 4.124ab (Nisvasatattvasamhita f 40r3) (> 
Svacchandatantra 7.308d): samanah samatam nayet, and Sardhatrisatikalottara 


Genesis and Development of Tantrism 

It comes as a surprise that Anandagarbha attributes the extreme Tantric 
practices that he details here to Vaisnavas, since nothing of this kind has been 
noted in their known literature. Because of this and because the use of female 
consorts, cow-flesh, urine and other products of the male and female body in the 
propitiation of deities for the attaining of supernatural powers or effects appears 
in our sources to be the hallmark of the Saiva Vidyapitha, and of the Picumata in 
particular, 544 it is tempting to propose that Anandagarbha has made a mistake 
and that had he been better informed or less careless he would have attributed 
these practices to those whom we know to have adopted them. But this cannot 
easily be accepted in the light of the fact that he backs up his attribution by 
citing a verse that supports it. I conclude, therefore, that his claim is rather evi- 
dence that some Vaisnavas had assimilated the transgressive, Sakta Saiva style 
of observance, just as the Buddhists had. In any case, whatever the accuracy of 
this attribution, it is extremely unlikely that Anandagarbha did not also have 
the Sakta Saivas in mind when he referred to "those devoted to the Tantras of 
Visnu and other [gods]". 

Similarly Sraddhakaravarman, one of the Indian teachers of the Ti- 
betan translator Rin chen bzang po (958-1055), says in his *Yoganiruttara- 

lO.lOcd: samam nayati gdtresu samano ndma mdrutah. 
544 See, e.g., Picumata f. 280v4: 67.71 saktigarte ksipel lingam tatah pujd<m> 
samdrabhet | gati-r-dgatiyogena saktiviksobhatatparah 'He should insert his pe- 
nis into the vagina of his consort and then begin the worship, intent on 
bringing his consort to orgasm through to-and-fro motion'; f. 106v3-4: 22.152 
saktim tu ksobhayen mantri vidydydstasatam japet | mantrasya vd japec caiva 
svaydgavidhicoditam || 153 dravyaprdsya purd krtva gomamsam kincisamyutam 
| surasthina samdyuktam pistam pindikrtan tatha || 154 ksobhadravyena 
sammardya lihgdkdram tu karayet | praksiped yonimadhye tu nimisam cdlya 
pidayet || 155 mantram uccdrayen mantri samkhydydstasatam tatha | karsayitvd 
tu tarn lingam gudikam karayet tatah || 156 japarcanavidhau nityam pujayet 
sddhakottamah 'The Mantra-adept should arouse his consort and [as he does so] 
repeat the Vidya 108 times. He should do the repetition of his Mantra as pre- 
scribed in the procedure for his set of deities. First he should swallow the sub- 
stances. Then he should grind cow-flesh mixed with faeces and surdsthi (urine?) 
into a ball, kneed it with the ejaculates, make it into the shape of a Liftga, insert 
it into [his consort's] vagina, move it about for a short while and then compress it. 
The Mantra-adept should utter the Mantra 108 times, then withdraw the Linga, 
and make it into a pellet. The best of Sadhakas should always offer [this] when he 
performs the repetition of the Mantras and the act of worship'; f. 10v5: gomamsam 
guggulam caiva pinyakam lasunam tatha || 3.210 siddhyartham gudika hy eta 
homayen nityakarmani | mandate tarpanam krtva gomamsasuraydnvitam 'Cow- 
flesh, bdellium, oil-cake, and garlic: he should offer this [mixture as a] pellet into 
the consecrated fire in his daily ritual'; f. 141v2 (28.38cd): gomamsam surayd 
misram homayita vicaksanah 'The adept should offer into the fire cow-flesh mixed 
with wine'; f. 39v3 (5.40ab): sampute sthdpayitvd tu mutrahomam tu karayet 'He 
should place urine in a bowl and offer it into the fire'. 


The Saiva Age 

tantrarthavatarasamgraha, referring to the Yogatantras as the Tantras of 
Method (Upayatantras) and to the Yoginitantras as the Tantras of Wisdom 
(Prajnatantras): 545 

A Method Tantra is one in which the Mandala shows mainly male deities in order 
to train (vini-) men and insiders (svayuthya-), whereas a Wisdom Tantra is one 
in which, in order to train women and non-Buddhist outsiders (bahyatlrthika-), 
the Mandala shows mainly female deities, deities, that is, who are appropriate 
for these. 546 A Method Tantra is one that exhibits deities that purify the outer 
and inner aggregates of personality (skandhah), the elements (dhatavah), and 
the faculties and their objects (dyatandni), whereas a Wisdom Tantra is one that 
exhibits deities that purify the outer and inner channels of the vital energy (nadi) 
and the Bodhicitta [semen]. A Method Tantra is one that exhibits deities [whose 
appearance and conduct are] in conformity with the [norms of] the world, whereas 
a Wisdom Tantra is one that exhibits deities [whose appearance is] contrary to 
[these norms of] the world. 

Since Sraddhakaravarman states here that the predominance of female deities 
is designed to recruit non-Buddhists he can mean only the followers of Sakta 
Saivism, since there is no other known group to whom this feature would have 
been particularly appealing. As for the other features that he identifies as dis- 
tinctive of the Yoginitantras, he does not state explicitly that they were intro- 
duced with the same purpose in mind; but it seems to me probable that he means 
this to be understood, since the transgressive character of these deities, his third 
distinctive feature, is indeed a fundamental characteristic of the goddesses wor- 
shipped by these outsiders. 

The Buddhism sponsored by the Palas had come a long way: too far, in fact, 
for those conservative Buddhist monks at Vajrasana who adhered to the ancient 

545 rNal 'byor bla na medpa'i rgyud kyi don la 'jugs pa bsdus pa, ff 103v7— 104r3: gang 
du skyes pa dang rang gi sde pa 'dul ba'i phyir lha po'i mam pa mang par ston pa'i 
dkyil 'khor ni thabs kyi rgyud do | gang du bud med dang phyi rol mu stegs can 'dul 
ba'i phyir de dag dang rjes su mthun pa'i lha mo'i mam pa mang pa'i dkyil 'khor 
ston pa ni shes rob kyi rgyud do \ gang du phyi nang gi phung po dang khams dang 
skye mched kyi mam par dag pa'i lha ston pa ni thabs kyi rgyud do | gang du phyi 
nang gi rtsa dang byang chub kyi sems mam par dag pa'i lha ston pa ni shes rab kyi 
rgyud do | gang du 'jig rten dang rjes su mthun pa'i lha'i mam pa ston pa ni thabs 
kyi rgyud do | gang du jig rten dang 'gal ba'i lha'i mam pa ston pa ni shes rab kyi 

546 p ar t f thjg formulation, namely the doctrine that the Yogatantras are designed to 
appeal to men and the Yoginitantras to women, has scriptural status, being found in 
the mKha' 'gro ma'i dra ba'i rdo rje gur rgyud (Dakinivajrapanjaratantra), f. 104v5— 
6: skyes bu mams ni gdul ba'i phyir \ mal 'byor rgyud ni yang dag bshad \ btsun mo 
mams ni bsdu ba'i phyir | mal 'byor ma yi rgyud bshad do 'The Yogatantras were 
taught in order to train (*vinayanaya) men. The Yoginitantras were taught in order 
to recruit (*samgrahdya) women'. 


Genesis and Development of Tantrism 

Buddhism of the Sravakayana. For according to the testimony of Taranatha 
they broke up the silver image of Heruka in the temple and burnt the collection 
of Tantras housed there, saying that these were the teachings not of the Buddha 
but of Mara, the evil obstructor of the Buddha's enlightenment. 547 

The Reflux of Buddhist Saktism into the Saktism of Bengal. In- 
deed, Buddhism had assimilated the Sakta Saiva style of religion so thoroughly 
that some of its creations went on to be adopted into the later Sakta Saivism of 
eastern India with little or no revision. This is the case with the goddesses Chin- 
namasta and Ugratara. The Buddhist origin of Chinnamasta is certain, since 
SVAHA, and the two companions that flank her are Dakini and Varnini. 548 In the 
Buddhist prototype the flanking goddesses are Vajravarnani and Vajravairocani, 
and the Mantra for recitation (japamantrah) is OM SARVABUDDHADAKINIYE 
SVAHA. 549 Moreover, the procedure of her visualization retains features dis- 
tinctive of her Buddhist Sadhana, notably that one is to visualize the goddess 
standing on a red sun-disk marked with a Yoni triangle on a white lotus in one's 
navel. 550 The only differences here are that in the Buddhist Sadhana the triangle 

547 Rgya gar chos 'byung, p. 168, 11. 14-: he ru ka'i sku dngul las byas pa chen po zhig 
dang | sngags kyi glegs bam mang dag cig yod pa si nga gling pa sogs nyan thos 
se ndha pa ga' zhig gis 'di dag ni bdud kyis byas pa'o zhes byas nas | glegs bam 
mams kyis bud shing byas | sku gzugs de yang dum bur bgos nas rnyed pa byas 
so 'There was a great silver statue of Heruka and many manuscripts of [texts of 
the] Mantra[naya]. Some Saindhava Sravakas from such [regions] as Sri Lanka, 
saying that these manuscripts had been created by Mara, used them as fuel, and, 
moreover, after dividing up the image into pieces pocketed them'; HBI, p. 279. 

548 Saktapramoda, p. 222 (her Mantra); pp. 221, 224-225 (the visualization of Chinna- 
masta, Dakini and Varnini) 

549 Abhisamayamanjari, pp. 151—152. 

550 Saktapramoda, pp. 224-225, Purascaryarnava, p. 816, Karmakanda, vol. 
4, p. 239d-240a (in the Kashmirian Saktasraddha): svanabhau nirajam 
dhyayec chuddham vikasitam sitam | tatpadmakosamadhye tu mandalam canda- 
rocisah | japakusumasamkasam raktabandhukasamnibham | rajahsattvatamo- 
rekhayonimandalamanditam | madhye tasya mahadevlm suryakotisamaprabham 
| chinnamastam kare vame dharayantlm svamastakam | prasaritamukhim 
bhimam lelihanagrajihvikam \ pibantim raudhirim dharam nijakanthavinirgatam 
| viklrnakesapasam ca nanapuspasamanvitam \ daksine ca kare kartrim 
mundamalavibhusitam | digambaram mahaghoram pratyalidhapade sthitam 
| asthimaladharam devlm nagayajhopavltinim | ratikamoparistham ca sada 
dhyayanti mantrinah 'He should visualize a pure, open, white lotus in his navel, 
the disc of the sun in the centre of the seed-pod of that lotus with the colour of 
the Japa flower, resembling the red Bandhuka blossom, adorned by a Yoni triangle 
with [three] linesf, red, white, and black representing the Gunas] Rajas, Sattva, 
and Tamas. At its centre Mantra adepts always visualize the Great Goddess Chin- 


The Saiva Age 

has the strictly Buddhist name dharmodayd and that the goddess is visualized 
as a transformation out of a yellow HRIH visualized in that triangle. 551 

In the case of Tara the Buddhist origin is even more apparent, since here the 
dependence extends to textual borrowing. For the Sakta literature of the worship 
of this goddess has incorporated the Mahacinakramatarasadhana of the Bud- 
dhist Sasvatavajra, which appears almost in its entirety in the eleventh chapter 
of the Sakta Phetkarinitantra. 

I am unable to determine within narrow limits how long after Sasvatavajra 
this Tantra was composed. 552 The earliest mention of the text in sources known 
to me is in 2.15 of the Sarvolldsatantra of Sarvanandanatha, in a list of a canon 
of sixty-four Tantras cited from the Todalatantra but not appearing in the pub- 
lished text of that work. It is probable that Sarvanandanatha, who wrote his 
work in Senhati in what is now Bangladesh, was born around the beginning 
of the fifteenth century. 553 It is tempting to assume that the PhetkdrinT was 
written at a time closer to Sasvatavajra's than to Sarvanandanatha's, that is to 

namasta shining like ten million suns, holding her own [severed] head in her left 
hand, fearsome, with the mouth [of her severed head] open wide, with the tip of 
her tongue licking greedily, drinking the stream of blood that gushes from her neck, 
her hair loosened, adorned with various flowers, holding a chopping-knife in her 
right hand, adorned with a garland of heads, naked, most terrible, standing in the 
Pratyalidha posture, with a necklace of bones and a snake as her sacred thread, 
standing on Kama and Rati'. 

551 Abhisamayamahjarl, p. 151: svandbhisthasuklakamalasuryasthitasindurdruna- 
dharmodaydmadhye pitahrihkdrajd svayam eva kartitasvamastakam vamahasta- 
sthitam dhdrayanti . . . 'Arising by transformation of a yellow syllable HRIH in the 
centre of a vermilion-red Dharmodaya triangle upon a sun[-disc] on a white lotus in 
his navel, holding her own head, which she herself has severed, in her hand . . . '. 

552 rp^g t a k e _ over f Sasvatavajra's Sddhana of Ugratara (= Sddhanamdld 101) by 
the Phetkdrinitantra and its subsequent influence have been demonstrated by 
Buhnemann (1996). Sasvatavajra flourished around the last decades the tenth 
century and the first decades of the eleventh. His Bdhyapujdvidhi (= Sddhanamdld 
252), Hastapujdvidhi (= Sddhanamdld 253), and Cakrasamvarabalividhi are found 
in the series of ritual texts published in FlNOT 1934 from a manuscript brought 
to China in 1057 by the Dhyana master Baocang on his return from India. His 
Sddhana of Ugratara is found in the *Sddhanasataka (a facsimile of an undated 
Sanskrit palm-leaf manuscript from Tibet has been published in Buhnemann 1994 
= T5h. 3306 ff.) and was translated into Tibetan by the Indian Pandita *Amogha- 
vajra and the Tibetan monk Bari Rin chen grags of Khams (Toh. 3373; DT, Rgyud, 
Mu, f. 49vl, colophon: rgya nag po'i rim pa'i sgrol ma'i sgrub thabs slob dpon rtag 
pa'i rdo rjes mdzad brjogs so | pa ndi ta don yod rdo rje dang khams pa lo tsd ba 
dge slong ba ris bsgyur cing zhus so). The latter was born in 1040 (Blue Annals, 
pp. 73 and 405) and was appointed to the chair of Sa skya in 1103 (Blue Annals, 
p. 211). A Sanskrit manuscript of his most important work, his commentary on the 
Laghusamvara, translated by Bu ston Rin chen grub (Toh. 1410), survives in the 
Potala Palace in Lhasa, where it awaits study. 

553 Sanderson 2007b, p. 236, fn.89. 


Genesis and Development of Tantrism 

say, when the Buddhist Mantranaya was still at its height in eastern India, be- 
fore the destruction of the great monasteries around 1200. But this destruction 
did not eliminate Tantric Buddhism and its literature from the region at a sin- 
gle stroke. For it was still alive in the early fifteenth century, when Vanaratna 
(1384-1468) travelled to Tibet in 1426, 1433, and 1453, gave various Tantric ini- 
tiations, notably in the Kdlacakra according to the system of Anupamaraksita, 
and assisted in the translation of Tantric texts, as is attested in the biography 
of this extraordinary figure given by Gzhon nu dpal (1392-1481), 554 who col- 
laborated with him in a translation of the *Trayodasdtmakasricakrasamvara- 
mandalavidhi (Toh. 1489). We also have the Vanaratnastotrasaptaka, a San- 
skrit hymn in praise of Vanaratna composed during his lifetime by a devout 
lay Buddhist Aditya, whom both the Sanskrit and Tibetan colophons say was 
a native of Magadha; 555 and we have a manuscript of the Mahayana classic Bod- 
hicarydvatdra copied by a lay Buddhist in Bengali characters at Venugrama in 
1436. 556 

After her incorporation from the Mantranaya Tara became with Daksinakali 
and Tripurasundari one of the three principal deities in the east-Indian Sakta 
system of the ten Mahavidyas, which soon became widely disseminated through- 
out the subcontinent. Thus in a passage cited from the scripture Jndnadvipa in 
the Sarvolldsatantra (3.1-29) the ten Mahavidyas are said to be [Daksina]kali 
(Syama), Tara, and Tripurasundari (Sodasi), with the third dividing into eight: 
herself and the seven others that make up the total often, namely Bhuvanesvari, 
Bhairavi, Chinnamasta, Dhumavati, Bagalamukhi, Matarigi, and Kamala. The 
centrality of these three goddesses is reflected in the corpus of east-Indian Sakta 
scriptures. The Todalatantra teaches the rites of these three alone, and the 
Brhannllatantra follows the same model but adds Kamakhya, the great goddess 
of Assam. Their centrality is also evident among the Paippaladin Atharvavedins 
of Orissa; for when they absorbed the influence of the Saktism of Bengal in the 
latest stratum of their diverse Angirasakalpa corpus it was principally the rites 
of Daksinakali and Tara that they adopted. 557 

The importance of Tara in late east-Indian Saktism is independently 

554 Blue Annals, pp. 797-805. On the career of Vanaratna see Erhard 2004. 

555 Hahn 1996, p. 37: samdptam idam [vanajratnastotrasaptakam | krtir magadha- 
deslyadityanam iti; p. 40: dpal Idan bla ma nags kyi rin chen bstod pa bdunpa 'di 
ni rdzogs so | yul ma ga dha nas byung ba'i bsnyen dam pa nyi ma pa zhes bya bas 
mdzad pa'o (*samaptam idam srlguruvanaratnastotrasaptakam | krtir magadha- 

556 Shastri 1917, p. 21: ASB MS 8067. The scribe identifies himself as Sadbauddha- 
karanakayasthathakkura Amitabha. 

557 Sanderson 2007b, pp. 235-236, fn. 88. 


The Saiva Age 

confirmed by the existence of substantial texts devoted exclusively to her 
worship, notably the Tararahasyavrtti of Gaudiya Sarikara composed in 1630, 
the Tarabhaktisudharnava, a work in some 11,000 verses composed by Nrsimha 
Thakkura c. 1688, the Tarabhaktitarahgini of Kasinatha, composed in 1682 
at the request of Krsnacandra, Maharaja of Nadia in West Bengal, and two 
other works with the same title, one by Vimalanandanatha and the other by 

The Jains' Adaptation of the Saiva Mantrasastra 

Jainism too enjoyed royal support during this period, notably in western 
India under the Caulukyas and in Karnataka among the Garigas of Talakad, 
the Rastrakutas, and Hoysalas; 558 and it too developed a Tantric ritual culture 
along Saiva lines for the propitiation (aradhana) of Mantra-goddesses for mun- 
dane benefits using Mudras, Japa, and offerings into fire (homah). Among god- 
desses worshipped in Jaina rites for such purposes are Laksmi and Vagisvari 
(Sarasvati) belonging to the higher world, the Vidyadevis belonging to the mid- 
dle, 559 and, most important, in the lower world the Yaksi attendants of the 
Tirtharikaras, associated with major Jaina pilgrimage sites, notably Ambika 
(/Kusmandini), the attendant of Neminatha at Girnar, Cakresvari, the attendant 
of Rsabha at Satrunjaya, Padmavati, the attendant of Parsvanatha at Sravana 
Belgola, and Jvalamalini, the attendant of Candraprabha. 560 

That these deities were developed on the basis of the Saiva tradition 
is more transparently obvious here than in Buddhism. Thus the Bhairava- 
padmavatikalpa, the Digambara Mallisena's Paddhati on the propitiation of 
Padmavati, written in 1057 equates her with Totala, Tvarita, Nitya, Tripura, 
and Tripurabhairavl, all well-known Mantra-goddesses of the Sakta Saivas. 561 

558 See Stein 1998, especially pp. 147-152. 

559 In the classical listing these are the following eighteen: Rohini, Prajnapti, 
Vajrasrhkhala, Vajrankusa, Apraticakra, Purusadatta, Kali, Mahakali, Gauri, 
Gandhari, Sarvastramahajvala, Manavi, Vairotya, Acchupta, Manasi, and 

560 For images of Ambika, Cakresvari, Padmavati, and Jvalamalini see, e.g., AIISPL, 
Accession numbers 45246, 10029, 58659, and 19995. On the cult of Padmavati see 
Jhavery 1944. On the cult of Jvalamalini see Settar 1969. 

561 On the worship of goddesses in Jainism and their division between the three worlds 
(urdhvalokah, tiryaglokah, and adholokah) see Cort 1987. On the centrality of 
the culture of Mantras and Mantrasiddhas in medieval Jainism see the survey 
and analysis by Paul Dundas (1998), who writes there of "the Jain mantrasastra' & 
partial linkage to an ultimately Saiva-inspired style of religiosity" (p. 36), of the 
Jhanarnava of the Digambara Subhacandra, probably in the tenth century, that 
it "blends much of the "software" of Saiva mantrasastra with specifically Jaina so- 


Genesis and Development of Tantrism 

Unlike Saivism, Pancaratra, and Tantric Buddhism in its mature form, 
Jaina Tantrism did not claim to offer Jainas a new path to liberation. It 
remained entirely focused on mundane benefits. Nonetheless it was not the 
preserve of the laity. Monks produced the manuals and monks were held to 
perform these propitiations. Thus Yasobhadrasuri and other Mantra-adepts 
(mantrikah) use the power that they have obtained by propitiating the goddess 
Kurukulla to unblock the throat of Devacarya when on the sixteenth day of a 
debate in the court of the Caulukya Siddharaja between him and the Digambara 
Kumudacandra the latter had used his supernatural power to silence him 
by causing him to choke; 562 the Jaina Guru of king Ajayapala undertakes a 
two-month propitiation of Ambika on the Raivataka mountain at Girnar in 
order to gain for himself the boon of equality with the renowned Svetambara 
Hemacandra and for his patron that of equality with Kumarapala, the great 
Caulukya king of Gujarat. 563 Hemacandra, Devendrasuri, and Malayagirisuri 
go to the same mountain at night to undertake the propitiation of the Siddha- 
cakramantra, after first performing preliminary rites to summon the presiding 
goddess Ambika into their presence; 564 and Hemacandra propitiates the spell- 
goddess Tribhuvanasvamini in Anahillapattana, the Caulukya capital, in order 
to ask her about the previous birth of his pupil Kumarapala. 565 

As in the non-Jaina tradition the goddesses were put to work to serve the 
interests of rulers. The Prabandhacintamani of Meruturigacarya, written at 
Vardhamana (Vadhvan) in eastern Kathiavad in 1304, claims that Padmavati 
was propitiated by means of a fire-sacrifice by a Digambara monk in order to 
protect Varanasi, the capital of king Jayacandra (in the late twelfth century), 
from attack by a Muslim army; 566 bards in Karnataka at the court of Yasodhara 

teriological concerns" (p. 35), and of the Bhairavapadmavatlkalpa that it "contains 
an account of the well-known six magical arts (satkarmani), not greatly dissimilar 
from their Hindu equivalents" (p. 33). 

862 Merutunga, Prabandhacintamani, p. 169: sodase dine akasmike devacaryasya ka- 
nthdvagrahe mantrikaih sriyasobhadrasuribhir atulyakurukulladeviprasadalab- 
dhavarais tatkanthapithat ksanat ksapanakakrtakarmananubhavat kesakandukah 
patayam cakre. 

863 Kumarapaladevaprabandha §54: cintitam devataradhanam vina manorathanam 
siddhir na | ato raivatake gatva devim ambam paritosya hemacaryasamo bhavi- 
syami | upavasatrayam tad anu talahattikdyam paranam | ekah paricaryakarah | 
evam masa 2 tapahprante devy amba pratyaksa jata karyam vada | tenoktam yadr- 
sah kumarapaladevas tadrsam ajayapaladevam yadrso hemacaryas tadrsam mam 

864 Kumarapalaprabodhaprabandha §61. On the worship of the Siddhacakra see 
Jhavery 1944, pp. 167-169. 

865 Kumarapaladevaprabandha §21. 
566 Prabandhacintamani, pp. 294-295. 


The Saiva Age 

are said to have invoked Aparajita to secure the king victory in battle; 567 and 
these powers are fully confirmed by the manuals for these rites. According to 
the unpublished Jvalamalinikalpa, composed by the Digambara Indranandin in 
Karnataka in 939, the benefits that can be attained by propitiating Jvalamalini 
include the splitting open of the gates of enemy forts; and the Bhairava- 
padmdvatlkalpa teaches a spell (vidyd) for making one's enemies fall asleep 
and magical receipts both for causing dissension among them (vidvesanam) and 
causing their death (mdranam). Moreover, Padmavati was the lineage goddess 
(kuladevT) of a number of Jaina ruling houses in Karnataka 568 and functioned in 
this capacity much as she would have done if they had not been converted. Thus 
she appears in a local manifestation as the Padmavati of Sasakapura (Sosavuru) 
in a Jaina myth of the origin of the name of the Hoysala (/Poysala) dynasty 
related in an inscription of 1133. 569 When a Jaina ascetic Yogin was trying to 
subjugate this goddess with a Mantra and a tiger sprang out to break its power 
the ascetic commanded king Sala, saying "Strike [it], O Sala" (poy sala). 570 The 
king then worshipped the goddess under the name Vasantika. Since this story 
introduces an account of the conquests of the dynasty it is probable that the 
goddess is seen here in the manner of the martial lineage goddesses of the Sakta 
Saiva type venerated by non-Jaina kings during the early medieval period as 
the source of their sovereignty and military might. 

In one important respect, however, Jaina lineage goddesses were bound to 
differ from their non-Jaina counterparts. Since Jainas are the strictest of vege- 
tarians and are rigorously opposed to the harming of any living creature, their 
goddesses, like those of the Buddhists, had to renounce the animal sacrifices that 
were so conspicuous a part of their cult in non-Jaina lineages. 571 Thus the Osval 

567 Cort 1987, p. 248. 

568 Notably the Silaharas, Rattas, and Santaras; see Cort 1987, p. 243. 

569 EC 5:124. 

670 Cf. EI 6:10, 1. 6: sa hoy saleti prdpat tarn kila vinihatya hoysaldkhydm. 

571 In the Buddhist case, however, animal sacrifice, though unusual, does occur. We 
see it in the mahdbali sacrifice performed by the Buddhist Newars at Lagankhel 
on the occasion of the chariot festival of Bugmalokesvara (Karunamaya); see Sin- 
clair 2008. Nor is this a recent innovation. See Catusplthatantra ff. 30r2-32r3. 
The Mantra for the Bali there (f. 31v2-) is derived from a Saiva prototype seen 
in the Vidyapitha's Nisisamcdra (14.56—63; ff. 47v5— 48v2: ekavrkse smasdne vd 
. . .). My pupil Peter-Daniel Szanto has kindly informed me (personal communi- 
cation, 4 March, 2009) that the verses that immediately precede that Mantra in 
this manuscript, containing the reference to sanguinary offerings, are not part of 
the original Catuspitha but have been added from the Catuspithamandalopdyikd 
of Caryavratipada (19.30-33 [f. 20rl). On that work, its author, and the incorpo- 
ration of material from it in this MS of the Catuspitha see Szanto 2008a. He 
has also drawn my attention to references to sanguinary offerings elsewhere in the 
Catuspitha itself, in the Sadhana of Dakinl (2.4.63-66) and in that of Cusini (2.4.75), 


Genesis and Development of Tantrism 

Jainas of Rajasthan and Saurashtra hold that their lineage deity Saccika or Sac- 
ciya adopted her present non-violence only when she and they were converted 
to Jainism by the monk Ratnaprabhasuri, probably in the twelfth century, 572 in 
consequence of his having miraculously cured a boy of snake-bite when he had 
already been thought dead and prepared for cremation. They claim that be- 
fore their conversion they had been Rajput warriors — a claim also found among 
other Jain castes — 573 and she a fierce Camunda whom they propitiated with the 
Tantric rites of the Vamamarga. Her pre-Jaina past is still visible in her tem- 
ple at Osian near Jodhpur, the Osvals' original home. For the outer wall of her 
innermost shrine shows images of Camunda, Mahisasuramardini, Sitala, and a 
naked Bhairava. 574 

We have another story of the conversion of a lineage goddess in Jaina 
accounts of the life of the Caulukya king Kumarapala of Gujarat (r. 1143- 
1174), who converted from Saivism to Jainism under the influence of the 
illustrious Svetambara scholar monk Hemacandra. According to these accounts 
Kanthesvari, the lineage goddess of the Caulukyas, and the other goddesses 
associated with her had always been placated during the nine days of the annual 

and to a reference to the attracting of animal and human victims (pasuh) at the end 
of the ninth chapter of the Vajradaka. That passage is derived from Laghusamvara 
32.1-2 and 31.2-3b. See also here p. 182, on human sacrifice. 

572 See Dundas 2002, p. 149. 

573 On the claims of Rajput ksatriya ancestry among the Jain castes of the Osvals, 
Khandelvals, Agravals, and Srimals see Babb 1993, pp. 7-8. 

574 Agrawala 1954 and 1956; Cort 1987, pp. 243-244; and Babb 1993, pp. 9-10, 
following accounts in Bhutoriya 1988. For photographs of the Camunda and 
Mahisasuramardini see AIISPL, Accession numbers 59386 and 59388. An account 
of the conversion of Saccika is found in a chronicle, the Upakesagacchapattavali, 
of the monastic community followed by the Osval laity, which ends with the in- 
stallation of Siddhasuri in [Vikrama] 1655. See pp. 237-238 of the translation by 
Hoernle (1890), who does not provide the original, for which see Agrawala 1954. 
Ratnaprabhasuri describes Saccika in that account as follows (Hoernle's transla- 
tion, p. 237), addressing her former devotees: 'O ye faithful, ye should not go to the 
temple of Sachchika-devi; she is merciless, and incessantly delights in hearing the 
sound of the breaking of bones and the killing of buffaloes, goats, and other animals; 
the floor of her temple is stained with blood, and it is hung about with festoons of 
fresh skins; the teachers of her devotion, rites, and service, are cruel men; she is 
altogether disgusting and horrible'. The text continues: 'Hearing these words of the 
Acharya, they replied, — "What you say, O Lord, is quite true; but if we do not go to 
worship that cruel Devi, she will slay us and our families." The Acharya, however, 
promised to protect them; whereupon they ceased to go any longer to the temple 
of the Devi'. Ratnaprabhasuri then goes on to convert the goddess, a tradition also 
asserted in an inscription of 1598 (Cort 1987, p. 244). Thereafter, it is said, she 
would accept no sanguinary offerings and not even red flowers, because they resem- 
ble such offerings. 


The Saiva Age 

Navaratra festival by the sacrifice of thousands of goats and buffaloes. 575 But 
this stops when Kumarapala, now a convert to Jainism, declares a fourteen-year 
ban on the taking of life. Kanthesvari appears before the king and demands to 
know why she and the other goddesses have been denied their usual sacrifices. 
When he explains that he cannot sacrifice to her now that he is a Jaina she is 
enraged and strikes him on the head with her trident, causing leprous sores to 
break out on his body. Hemacandra miraculously cures his affliction, tries to 
persuade the goddess to accept in future offerings of vegetarian food of equal 
value, and when this fails binds her with a Mantra. Thoroughly humbled, she 
begs the king to free her, promising that if she is released she will give up her 
ways and work instead to police his ban on the slaughter of animals throughout 
his realm. With Hemacandra's permission he releases her and she takes to her 
new role as the king's informer with all the zeal of the convert. 576 She reports a 
vassal king in Saurastra for secretly butchering goats in his home: Kumarapala 
sends his minister Udayana at the head of an army to punish him. 577 She 
reports a merchant for plucking a louse from his wife's head and crushing it: 
his entire property is seized and the money used to fund the building of a Jaina 
monastery, named accordingly the Monastery of the Louse (Yukavihara). 578 

575 Three thousand seven hundred goats and thirty-seven buffaloes were to be sac- 
rificed: a hundred goats and one buffalo on the first day, two hundred goats 
and two buffaloes on the second, three hundred goats and three buffaloes on the 
third, and so on, so that nine hundred goats and nine buffaloes were sacrificed 
on the ninth (Mahanavami). See Somatilakasuri, Kumarapaladevacarita vv. 387- 
389: suddhasamyaktvaputatma mahanavamiparvani \ kumarapalabhupala 
amigadibhir akhyata || 388 devi * kanthesvari (corr. :kamtesvariF,d.) gotradevi svam 
bhavyam ihate | ekam chagasatam caiko mahisah pratipaddine || 389 etavad eva 
dvigunam dvitiye divase punah \ trtiye trigunam yavan navame *navasamgunam 
(corr. :nava samgunam Ed.); and Kumarapalaprabodhaprabandha §75: athamarim 
pravartayati rajani asvinasuklapakso 'gat \ tatra *kanthesvaryadidevatanam 
(kanthesvaryadi corr. : kantesvaryadi Ed.) arcakair vijhaptam deva saptamyam 
sapta satani pasavah sapta mahisas ca devatanam puro diyante rajha | evam 
astamyam astau satani navamyam nava sataniti. In the editions of the 
Kumarapaladevacarita and the Kumarapalaprabodhaprabandha the goddess' 
name appears in the form Kantesvari. I have corrected this to Kanthesvari on the 
dubious strength of a passage in the Prabandhacintamani of Merutunga in which 
the author implies that she owes her name to the fact that in the eighth century 
Vanaraja, the founder of the Capotkata dynasty that preceded the Caulukyas at 
Anahillapattana, had a shrine built for her in the kanthah ('narrow entrance'?) 
of his palace (p. 35: tatha ca tena dhavalagrhakanthe kanthesvariprasadas ca 

576 Kumarapaladevacarita, vv. 387-396 and Kumarapalaprabodhaprabandha §75. 

577 Kumarapalaprabodhaprabandha §85. 

578 Kumarapaladevacarita, vv. 404-406; cf. Kumarapalaprabodhaprabandha §77. The 
same sources relate another occasion on which the Jaina Mantravada was used 
to curb a sanguinary goddess. Hemacandra and Yasascandra fly through the 


Genesis and Development of Tantrism 

Thus, while drawing heavily on the Sakta Saiva tradition of the propitia- 
tion of Mantra-goddesses, the Jain Mantravada, was bound to keep itself free of 
the sanguinary aspects of those cults and, also, one would assume, of all other 
transgressive elements that would conflict with the ascetic character of the Jaina 
path, notably the use of flesh and alcohol, and the employment of female consorts. 
However, that exclusion was not as complete as one would expect in respect of 
the last of these elements. This is apparent in the accounts of two of the pro- 
pitiations mentioned above. We are told that when Hemacandra, Devendrasuri, 
and Malay agirisuri undertook the propitiation of the Siddhacakramantra on the 
Raivataka mountain they did so with a Padmini in the person of the wife of a vil- 
lage headman as their Tantric assistant (uttarasadhakatvena). 579 How the wife 
of the village headman assisted in the propitiation is not stated. But the story of 
Hemacandra's propitiation of Tribhuvanasvamini is more explicit. Again he has 
the assistance of a Padmini. The daughter-in-law of a farmer is brought to the 
city for this purpose and the goddess shows her favour after Hemacandra has 

air from Anahillapattana to Bhrgupura (Bhrgukaccha, Bharukaccha, modern 
Bharuch/Broach) and attempt to tame the Tantric goddess Saindhava, who had 
possessed the minister Ambada. She shows her contempt for Hemacandra by stick- 
ing out her tongue. Yasascandra punishes her by pounding some grains of rice in 
a mortar. The first blow causes her temple to quake, and the second and third 
cause her image to shudder and then be dislodged. She falls at Hemacandra's feet 
begging for his protection. See Somatilakasuri's Kumdrapdladevacarita, w. 76-85 
and Kumarapalaprabodhaprabandha §87. Saindhava is no doubt the Sindhavai 
Ma whose temple is located outside the walls of Broach to the north, not far from 
the temple of Nilakantha. She was receiving goat sacrifices on Mahanavami up to 
the 1940s (Desai 1993, p. 48). According to Somatilakasuri, she was the principal 
of the non-Jaina deities of the city. Sindhavai Ma also has temples in Ahmedabad, 
near Bilimora, and Kayavarohana, Vadodara. 
579 Kumarapalaprabodhaprabandha §61: te ca trayah krtapurvakrtyah srl-ambika- 
krtasannidhyah subhadhyanadhiradhiyah sriraivatadevatadrstau triyaminyam d- 
hvdndvagunthanamudrakaranamantranydsavisarjanddibhir upacarair guruktavi- 
dhina samipasthapadministrikrtottarasadhakakriyah srisiddhacakramantram *a- 
sadhayan (em. : asadhayat Ed.). And those three, after performing the prelimi- 
nary service (purvaseva) and bringing about the presence of Ambika, with their 
mind firmly concentrated in the 'pure' mode of meditation, in the sight of the 
goddess of the Raivataka mountain, performed at night the Sadhana of the Sid- 
dhacakramantra following the procedure taught by the Guru, with all the [re- 
quired] rites of summoning, enclosing, making the Mudras, installing the Mantras 
[on their bodies], dismissing and the rest, with the actions of the Tantric as- 
sistant performed by that Padmini beside them'. According to the erotologi- 
cal literature Padmims are one of four classes of ideal love-partner (nayika); 
see, e.g., Pancasayakamanjari 1.6: sampurnendumukhi kurahganayana pinastani 
daksina mrdvahgi vikacaravindasurabhih syamatha gauradyutih \ alpahararata 
vilasakusala hamsasvana sadgatir lajjalur gurudevapujanapara syan nayika pad- 
mini; and in Tantric literature Hevajratantra 2.7.2-5 and Samvarodayatantra 


The Saiva Age 

repeated the Mantra for three days on the Padmini's vulva (tasyd yonau). 580 The 
text tells us that Hemacandra's mind remained undisturbed during this prac- 
tice, no doubt wishing to stress that he was not compromising the monastic rule 
of celibacy. Indeed there is no evidence of which I am aware that the Jaina 
Mantravada, unlike Saivism and Tantric Buddhism in its later phases, created 
two levels of discipline, one for ordinary practitioners and one for an elite that 
transcended the rules that apply to the first. Nonetheless, we see from this story 
that it had gone surprisingly far in this direction, too far for some, one suspects, 
who would have preferred monks to avoid any practice in which they could be 
suspected of departing from the straight and narrow Jain path of purification. 

Saivism in the Brahmanical Substrate 

As for the long-established brahmanical tradition, the Saivas saw it as sub- 
sumed within their own, accepting it as the only valid source of authority in what 
they saw as the lesser domain of mundane religion (laukiko dharmah). This per- 
ception is much emphasised in their literature, 581 and it is expressed through the 

580 Kumar apdladevaprabandha §21: atha srihemdcdryais tribhuvanasvdminim 
vidydm drddhayitukdmd bhdnddgdrikam kapardinam prdhur yan mehatdgrdme 
trihunasimhah kautumbikah | tasya putrds catvdrah \ laghor vadhuh padmini \ 
yadi sdydti tadd *tasyd avdcyapradese (corr. : tasy avdcyapradese Ed.) dinatrayam 
jape datte devT prasidati | etad atiduskaram | kapardinoktam | cintd na vidheyd | 
bhdnddgdrikas tatra gatah kautumbikagrhe \ tena satkrtah | prayojanam prstah | 
bhdnddgdrikenoktam laghuputravadhum mamdrpaya | tenoktam kim idam ddisasi 
| evam eva \ vicdro 'pi na kartavyah | tenoktam yadi bhavatdm *vicdre samdydtam 
idam (?) tadaivam astu \ sukhdsane 'dhiropya pattane samdgatah \ srihemasuribhih 
paramdnndhdraparair avikrtacittais tasyd yonau dinatrayam jdpah krtah | devT 
tustd 'Then Hemacarya, desiring to propitiate the spell-deity Tribhuvanasvamim 
said to his treasurer Kapardin: "There is a farmer called Trihunasimha in Mehata 
village. He has four sons. The wife of the youngest is a Padmini. If she comes here 
and I offer Japa for three days on her unmentionable part the goddess will favour 
me. This is extremely difficult [to accomplish]". Kapardin told him not to worry. So 
the treasurer went to the home of the farmer in that [village] and after being hon- 
oured was asked his purpose. The treasurer said: "Give me the wife of your youngest 
son". [The farmer] said: "Is this an order?". He replied that it was but that he should 
not be concerned. [The farmer] said: "So be it, if this is *what you have decided after 
due deliberation (?)". So [the treasurer] put her in a comfortable sedan and returned 
with her to the capital. The venerable Hemasuri did the Mantra-recitation on her 
vulva for three days, intent on eating paramdnnam, with his mind undisturbed [by 
lust]'. The goddess was pleased'. The food paramdnnam is, I presume, the dish of 
rice, milk, and sugar or jaggery otherwise known as pdyasam and considered the 
ideal food for offering to a vegetarian deity. 

581 It is encapsulated in the often cited words of their scripture Bhdrgavottara: iti 
varndsramdcdrdn manasdpi na lahghayet \ yo yasminn dsrame tisthan diksitah 
sivasdsane \ sa tasminn eva samtisthec chivadharmam ca pdlayet 'So he should not 
transgress the practices of his caste and [brahmanical] discipline even in thought. 


Genesis and Development of Tantrism 

collocation of the epithets paramamdhesvarah and paramabrahmanyah that is 
sometimes found with the titles of our kings in inscriptions. 582 

But the brahmanical tradition was not merely accepted by the Saivas. It 
was also influenced by them. During this period we find an ever-growing cor- 
pus of traditions that while claiming to be on the brahmanical side of the divide 
derive from the Saiva, both Saiva devotional literature assigned to the Puranas 
and a form of worship that followed Saiva models. In Puranic texts such as 
the Uttarabhdga of the Lifigapurdna, 583 the Kdlikdpurdna, the Devlpurdna, and 
the Agnipurdna, b8A the boundary between the Smarta and Tantric domains has 
almost completely dissolved, prompting the conservative brahmanical author 
Ballalasena, the twelfth-century Sena king of Gauda, to reject them as invalid 
as sources of the knowledge of religious duty, objecting particularly to their con- 
taining instruction on such matters as Saiva initiation and idol consecration. 585 

In reality there was no reasonable hope of turning the tide by this period, 
as had to be conceded even by so conservative an authority as the Nibandha 
on the Ydjnavalkyasmrti compiled by or under Aparaditya, the Silahara king 
of Korikana in the last quarter of the twelfth century. While firmly denying in 
general the validity of the practices taught in the Saiva scriptures, it admits a 
partial exception in the case of the Sthapaka, the priest who consecrates idols 
and shrines. It is admitted that he may draw on these texts to supplement the 

He should remain in the discipline in which he was when he was initiated into the 
Saiva religion and [at the same time] maintain the ordinances of Siva'; see SANDER- 
SON 1988, p. 662 (= 1990, p. 139); 1995, p. 23; 2005a, p. 389; 2007a, pp. 231-232. 
The Saivas' understanding of how the relation between the general, Vaidika ordi- 
nances and those of the Saiva scriptures should be perceived is explored at length 
in Sanderson forthcoming b. 

582 ^y e gee ^g com bination in the case of the Panduvamsins/Pandavas of Mekala in 
the fifth century (Shastri 1995, nos. II: I— II), the Sailodbhava Madhavaraja of 
Kongoda in the seventh (EI 6:14), the Pallavas Paramesvaravarman I (c. 669-690) 
and Narasimhavarman II (c. 690-728/9) (Mahalingam 1988, nos. 45, 53) around 
the turn of the seventh and eighth, the Bhanja Nettabhanja of Orissa in the eighth 
(EI 28:41, 11. 16-17), the descendants of King Nimbara of Kartikeyapura in Hi- 
machal Pradesh in the ninth and tenth (EI 31:38), and the Eastern Calukyas in 
the eleventh (EI 6:35; EI 6:36). 

583 On the presence of the Saiva Mantramarga in its Saiddhantika, Daksina 
(Bhairava), and Sakta forms in the Uttarabhaga of the Lifigapurdna see SANDER- 
SON 2005b, pp. 235-236. 

584 On the Agnipurdna's incorporation of the Saiddhantika Saiva Paddhati of 
Somasambhu see p. 65 above. 

685 In w. 55-67 of the introduction to his Ddnasdgara Ballalasena rejects on these 
and allied grounds the Garudapurdna, the Brahmapurdna, the Agnipurdna, the 
Vaisnavapurdna in twenty-three thousand verses, the Lifigapurdna in six thou- 
sand, the Devlpurdna, and parts of the Bhavisyapurdna. That he did not include 
the Kdlikdpurdna in his list strongly suggests that it postdates him. 


The Saiva Age 

ritual of consecration when installing a Siva, and likewise on the other appropri- 
ate bodies of non-Vedic scripture when consecrating images of the Goddess and 
the like, provided that his Vedic procedure needs to be supplemented, provided 
that the imported auxiliary does not offend the Vedic procedure in any way, and 
provided that he does not take the initiations (diksa) which those scriptures re- 
quire. 586 In other words it had to be conceded that a hybrid of Tantric and Vedic 
rituals procedures was already an institutional reality; and that this was so is 
confirmed by a Saiva source, which protests against their existence, insisting 
that patrons should engage only initiated Saiva officiants of full conviction, who 
would perform Saiva rituals of consecration uncontaminated by such hybridiza- 
tion. 587 

586 This position is established at length in the course of the commentary on 
Ydjnavalkyasmrti 1.7, which lists the valid sources of knowledge of religious 
duty (dharmamulam), namely Sruti, Smrti, and observation of the practice 
of exemplary brahmins, supplemented by personal judgement and preference 
where the other sources of knowledge leave scope for them. Aparaditya consid- 
ers at length and rejects the proposition that the scriptures of the Pasupatas, 
Saivas, Pancaratrikas, and others not rooted in the Veda (vedamula-) should 
be added to the list (vol. 1, p. 10, 1. 6 ff.). He concludes: tatas ca devapujadau 
narasimhapuranadiprasiddhaivetikartavya grahya nanya | evam dlksayam apy 
avagantavyam | na hi purdnaprasiddhayam dlksayam jatisodhanam asti (vol. 1, 
p. 14, 11. 17-19) . . . evam pratisthayam api puranadyuktaivetikartavyata grahya 
nanya tesam eva vyamisradharmapramanatvena bhavisyatpurane parijnatatvat 
(p. 15, 11. 1-2) 'And so the procedure for such [ritualsl as the worship of deities 
that may be adopted is that taught in such Puranas as the Narasimha- , and no 
other. The same should be understood to apply in the case of initiation. For in 
the initiation established in the Puranas the [objectionable Saiva] rite of the elim- 
ination of [the initiand's] caste is lacking. . . . Equally, in the case of rituals for 
the installation [of the image of a deity and the like only the procedure taught in 
Puranas and [related texts] may be adopted, since the Bhavisyatpurana acknowl- 
edges none but these as sources of valid knowedge of hybrid religious duty'. By 
'hybrid' (vyamisra-) Aparaditya means procedures that incorporate auxiliary ele- 
ments from the Tantras. The issue of this hybrid installation rituals is taken up in 
detail on pp. 16, 1. 1-19, 1. 12. 

587 This source is the Saiddhantika scripture Devyamata. It devotes several verses 
to distinguishing types of Sthapaka and to exhorting patrons to avoid all but one, 
who is described as learned both in the general Saiva scriptures and in the special- 
ized Tantras of Installation, as content with the teaching of Siva, focused wholely 
upon it, strictly adhering to the discipline of the initiated (samayacarah), with- 
out any inclination towards the scriptures of the uninitiated ipasusastram), tak- 
ing no pleasure in the mundane religion, but delighting in the religion of Siva 
alone: (2.16cd, 17ab, 19ab, 20ab): acaryah sivasastrajhahpratisthatantraparagah || 
... 17 sivasastrarthasamtustah samayacarapalakah | . . . 19ab sivasastraikacittatma 
pasusastraparahmukhah | ...20 virakto laukike dharme sivadharmanurahjitah. 
Sthapakas to be avoided are those who are Vaidika in their religious commit- 
ment and learning. Some of these have no more than a partial knowledge of the 
Tantras of Installation; but they should be avoided even if they mastered both 
the Tantras of Installation and the general Saiva scriptures (2.7-8b and 2.13-14): 


Genesis and Development of Tantrism 

Instances of incorporation of Saiva ritual in the Smarta domain can be ad- 
duced from most regions and periods; 588 but perhaps the most striking because it 
was so widely disseminated and accepted by those who considered themselves to 
be on the Smarta side of the divide is represented by the Prapancasara attributed 
to Sarikaracarya and the closely related Sdraddtilaka of Laksmanadesika. These 
two texts, which, I have argued, were composed in Orissa or on the basis of Oris- 
san tradition, most probably in the twelfth century, 589 present a system of ritual 
that differs from the properly Tantric only in its catholic character — in Smarta 
fashion it includes rituals of propitiation for all the main deities — , its avoidance 
of all the elements of 'impure' practice that the Smartas castigated in the Saiva 
cults of Bhairava and the Goddess, and its expurgation of doctrines that were 
contrary to what could be found in acceptably brahmanical sources, notably the 
doctrine of the thirty-six levels of reality (tattvani). 


Saivism, then, was undoutedly the most successful among the religious sys- 
tems that received royal patronage during the early medieval period. It was 
the most commonly adopted. Of the others some were absorbed by it and the 
rest while flourishing independently beside it came to remodel themselves along 
Saiva lines. 

No doubt there were many factors that led to Saivism's rise to dominance 
within this complex environment, and no doubt many of these will remain in- 
visible to us, since they could be discerned and weighed only if we had access 
to much more detailed evidence of the activities and motivations of individuals 
and institutions, both religious and political. Nonetheless, I venture a general 

The Early Medieval Process 

On the basis of the epigraphical record of acts of patronage, and consider- 
ing evidence of changes over time within the Saivas' prescriptive literature, I 

pratisthatantrakincijjnah pasusastranuranjitah \ tattvopadesahinas ca nacaryo na 
ca sadhakah || 8 tena samsthapitam lihgam siddhidam na kadd cana | ... 13 pa- 
davakyapramanajho brahmano vedaparagah | pratisthatantrakincijjnah sthapako 
na prasasyate || 2.14 pratisthatantratattvajnah sivasastravisaradah | so 'pi na stha- 
pakair istah pasusastranuranjitah. 

588 One of these, the assimilation of Sakta Saiva propitiation rites by the Athar- 
vavedic tradition of the Paippaladins of Orissa, has been demonstrated at length 
in Sanderson 2007b. 

589 Sanderson 2007b, pp. 230-233. 


The Saiva Age 

propose that the fundamental reason for the religion's success, underlying and 
structuring the mass of particulars now lost to view, was that it greatly increased 
its appeal to royal patrons by extending and adapting its repertoire to contain a 
body of rituals and theory that legitimated, empowered, or promoted key ele- 
ments of the social, political and economic process that characterizes the early 
medieval period. 

These elements were: 

1. the spread of the monarchical model of government through the emergence 
of numerous new dynasties at subregional, regional, and supraregional 

2. the multiplication of land-owning temples, both royal temples in nuclear 
areas and lesser temples in peripheral zones, often established by subor- 
dinate local lords, thus promoting the rural economy and the progressive 
penetration of the authority of the centre into new territories; 

3. the proliferation of new urban centres, both commercial centres that grew 
from below through a process of agglomeration, and planned settlements, 
growths from above, founded by rulers; 

4. the expansion of the agrarian base through the creation of villages, land 
reclamation, and the construction of water-reservoirs, wells, and other 
means of irrigation, with the steady growth in population that these de- 
velopments imply; and 

5. the cultural and religious assimilation of the growing population of com- 
munities caught up in this expansion. 590 

At the same time it took steps to integrate itself with the brahmanical sub- 

590 p or thjg positive characterization of the period I am indebted to the work of a num- 
ber of historians who in recent decades have shown the invalidity of the widespread 
view that it was a time of decline, de-urbanization, fragmentation, and general im- 
poverishment in the aftermath of a glorious classical age that culminated under the 
Gupta kings and ended with their demise. I acknowledge in particular the research, 
conclusions, and hypotheses of Noboru Karashima (1984), R. Champakalakshmi 
(1986), Hermann Kulke (1990, 1995a, b), Brajadulal Chattopadhyaya (1994), 
Upinder Singh (1994), Burton Stein (1994, 1998), James Heitzman (1995), and 
Cynthia Talbot (2001). That judgement, which owes more, one suspects, to the 
concept of the European Dark Ages after the collapse of the Roman empire than to 
unbiased analysis of India's epigraphical and archaeological record, has its coun- 
terpart in the not uncommon assessment that these centuries also witnessed a pro- 
gressive degeneration of Sanskritic literary, intellectual, and religious culture. It 
is refreshing to see that the work of those historians who are engaging vigorously 
with the epigraphical and archaeological evidence of the age has brought forth a 
view that is more consonant with the abundant literary evidence of intellectual and 
aesthetic vigour. 


Genesis and Development of Tantrism 

strate in ways that rendered it accessible and acceptable to a far wider con- 
stituency and therefore all the more appealing to rulers in their role as the 
guardians of the brahmanical social order. 

Saivism and Monarchy 

Saivism's engagement with the first and most crucial of these elements is 
apparent in the fact that from the seventh century onwards inscriptions and pre- 
scriptive religious texts reveal that Saiva brahmin Gurus were holding the posi- 
tion of royal preceptor (rajaguruh) in numerous new kingdoms both on the Indian 
subcontinent and in Southeast Asia and in this capacity empowering and legiti- 
mating the monarch's rule by granting him Saiva initiation (sivamandaladiksa). 
It might be thought that this would have been an unappealing step for any but 
the most reclusive and ineffectual of kings, since after initiation Saivas were 
obliged to adhere to a complex and time-consuming program of daily and oc- 
casional rituals. However, early in the development of the Mantramarga, the 
Saivas, no doubt in order to extend their recruitment and hence their influence, 
admitted a category of initiates who in consideration of the fact that they were in- 
capable of taking on these onerous duties were exonerated from doing so. 591 The 
king was considered to qualify for this less arduous route to liberation by reason 
of his royal obligations. He was therefore required to adhere only to the obli- 
gations of an uninitiated devotee of Siva taught in the texts of the Sivadharma 
corpus, which in his case were principally to support the religion and its institu- 
tions and to sponsor and appear in conspicuous ceremonies in the civic domain. 

Moreover, according to prescriptive sources the king's initiation was to be 
followed by a Saiva modification of the brahmanical royal consecration ceremony 
(rajyabhisekah), bestowed both on the king and his chief consort, and also given 
to the heir apparent at the time that he was consecrated to succeed to his father's 

591 The distinction between these two categories of initiate, those who receive initi- 
ation with post-initiatory duties (sabija diksa 'initiation with seed' ) and those 
who receive it without (nirbija diksa 'initiation without seed'), is not present in 
the earliest Saiddhantika scriptures, namely the corpus of Nisvasa texts found in 
the Nisvasatattvasamhita codex, the earliest of which, the Mulasutra, was probably 
composed at some time between 450 and 550, for which dating see the conclusions of 
a recent workshop on this text summarized in the newsletter of the Nepal-German 
Manuscript Cataloguing Project (Goodall and Isaacson 2007). On the relatively 
archaic character of the Nisvasa corpus see SANDERSON 2001, pp. 22-31 (archaic 
features listed in fn. 32, pp. 29-31), and Sanderson 2006. The category of exon- 
erated initiates appears later in the Kirana, the Paramesvara, and the Svaccha- 
nda, and, following the latter, in the Paddhatis. The textual evidence is given in 
Sanderson forthcoming a. 


The Saiva Age 

throne (yuvarajabhisekah). 592 

This new ceremony was added to the purely Saiva consecrations recognized 
by the core tradition, through which a Saiva Guru empowered an initiate to take 
office as a Sadhaka (sadhakabhisekah), a specialist in Mantra-rituals for super- 
natural effects (siddhih), and that through which a retiring Guru (acaryah) con- 
secrated his chosen successor (deary abhisekah), passing on to him his duties. In 
this way the monarch was incorporated as a third kind of Saiva initiate, who 
differed from the Sadhaka and the Guru not in the character of the initiation 
itself but in the consecration ceremony that followed it: while they were to be 
consecrated for purely Saiva functions, the king was to be consecrated to take 
up office as the "head of [the brahmanical social order of] the caste-classes and 
religious disciplines" (varnasramaguruh), 593 the role already assigned to him by 
brahmanical prescription. 594 

As the function of the Saiva consecration is modified in this case, so its form, 
though in general Saiva, incorporates distinctive non-Saiva elements appropri- 
ate to its mundane and brahmanical aspects, such as the inclusion of the royal 

592 The textual and epigraphical evidence for the practice of royal initiation, and the 
textual evidence for the king's exoneration from Saiva duties, and this ancillary 
Saiva modification of the brahmanical royal consecration ceremony are presented in 
Sanderson forthcoming a. On the brahmanical consecrations of the king, queen, 
and heir apparent see Sanderson 2005a, p. 382 and notes 115-117. 

593 Naimittikakarmdnusamdhdna f. 74vl: [4.118] varndndm dsramdndm ca gu- 
rubhdvdya bhupateh | yo 'bhisekavidhih so 'pi procyate diksitdtmanah 'I shall also 
teach the rite of consecration as the means by which a king, provided that he has 
received [Saiva] initiation, becomes the patron of the caste-classes and brahmanical 

594 Manusmrti 7.35cd: varndndm dsramdndm ca raja srsto 'bhiraksitd 'The king 
has been created as the guardian of the castes and disciplines'; Brhaspatismrti 
1.9ab: tasmdd varndsramdndm tu netdsau nirmitah purd 'he was created of old 
as the leader of the castes and disciplines'; Visnusmrti 3.1-3: atha rdjadharmdh. 
prajdparipdlanam | varndsramdndm sve sve dharme vyavasthdpanam 'Next the du- 
ties of the king: protection of his subjects [and] ensuring that the castes and [follow- 
ers of the] disciplines keep to their respective duties'; Visnudharmottara 2.65.55: 
varndsramavyavasthd tu tathd kdryd visesatah | svadharmapracyutdn raja sva- 
dharme viniyojayet 'And his special duty is to establish the castes and disciplines. 
The king must force those who have fallen away from their duties [as members and 
followers of these] to practice them'. The characterization of the king in accordance 
with these injunctions as the Guru of the castes and disciplines (varndsramaguruh) 
is a commonplace in our period. See, for example, Sdtvatasamhitd 24.16-17 (> 
Isvarasamhitd 17.14-15); Somadeva, Kathdsaritsdgara 12.6.85; Candraprabhasuri, 
Prabhdvakacarita v. 284ab; Ksemendra, Avaddnakalpalatd 2.60c and 27.22b. See 
also the cognate expressions sarvdsramaguruh and dsramindm guruh in Ne- 
tratantra 19.87 and 20.55b, varndsramadharmamarydddedryah and akhildsrama- 
guruh in Agamadambara , Act 2, prose after 20 and Act 3, v. 4, and varnaguruh in 
RdjatarahginT 3.85ab. 


Genesis and Development of Tantrism 

banners, weapons, and armour in the objects of worship, 595 the seating of the 
king on a platform covered with the skins of a fighting bull and a cat, 596 the 

595 Naimittikakarmanusamdhana, f. 75r4— vl (4.129c— 132): ghatesv abhyarcya lokesdn 
sdstrdn indrapura<h>sardn || 130 sivam agnih ca hetis ca ketum (conj. \ketus Cod.) 
cesddivedisu [Marginal glosses: khadgddi on hetis and dhvajacihnam on ketus] 
| samnidhlkrtya samtarpya pujayec cakravartinah || 131 udagvedisira<h>sthesu 
kalasesuktalaksmasu | anantddisikhandy*antdn (corr. : antd Cod.) digvidiksu 
yathdkramam || 132 tasyds tadvad adha<h>sthesu rudramdtrgandrthaddn \ 
grahdsurapalds*dkhydn (conj. :dkhya Cod.) bhogindm adhipdn api 'He should wor- 
ship Indra and the other Lokapalas together with their weapons in the vases, and 
then Siva, Agni, the [royal] weapons, and the [royal] banner on the altars begin- 
ning [with that] in the northeast. He should then summon, gratify, and worship 
the [eight] Universal Monarchs [, i.e. the Vidyesvaras], beginning with Ananta 
and ending with Sikhandin, in the vases whose required characteristics have been 
stated above, set on the northern altar, and likewise, below that [altar], the Rudras, 
the Matrs, Kubera, the Grahas, the Asuras, the flesh-eating [Raksasas], and the 
Naga lords'; f. 76r2-4 (4.141—142): sivdgnihetiketundm kdritdbhydm athdrcanam \ 
pahcagavyam carum tdbhydm dattvd ca dvijasodhanam || svdpayitvd tu tau tatra 
saraksau vedikddvaye | prtha<k> prdksirasau mahydm samyatau ksaumasayyayoh 
'He should make both [the king and queen] offer worship to Siva, the Fire, the 
[royal] weapons, and the [royal] banner, and then give them the five products of the 
cow, rice porridge [prepared on the sacred fire], and a tooth-cleaning twig. He should 
then have them sleep on the ground with their heads to the east on beds of linen 
on the surface of the two altars, having provided them with protection (saraksau). 
They should observe chastity [throughout the night].' For the protection mentioned 
here see the rites such those of protecting the beds by reciting of the Weapon-Mantra 
over them and surrounding them with Weapon-empowered lines of mustard-seeds, 
sesame-seeds, and ash set out in Uttarakdmika 23.54-59 (elaborating the related 
expression saraksdn svdpayen nisi) and Mrgendra, Kriydpdda 7.98c-103, both cited 
in Brunner 1977, pp. 216-221. As for the requirement that the king and queen 
should sleep with their heads to the east, this too expresses the relatively mun- 
dane nature of this consecration. For at this point in Saiva initiation ritual can- 
didates are to sleep with their heads to the east if they seek benefits other than 
liberation; see Mrgendra , Kriydpdda 7.99ab: bubhoksoh sayanam kurydd guruh 

596 Naimittikakarmanusamdhana f. 76v4— 5 (4.150— 152b): hetin astrena ketums 
ca varmand kahkatdny api [Marginal gloss on kahkatdni: samnahydni] | 
sugandhapuspadhupddyair naivedydntaih prapujya ca \\ anantddims ca *vidyesdn 
udagvedydm (conj. : ved+ + + + + vedyds Cod.) ca piirvavat | rudrddlms ca ghatesv 
istvd vedyor urdhvam athdstaret \\ brhaduksno 'tisurasya vrsadamsasya carfmani] 
After worshipping with offerings beginning with fragrant flowers and incense and 
ending with cooked food the weapons and the banners with the Weapon-Mantra 
and the cuirasses with the Armour-Mantra, he should worship Ananta and the 
other *Vidyesvaras on the northern altar (conj.) as before and after worship- 
ping the Rudras[, the Matrs, Kubera, the Grahas, the Asuras, the flesh-eating 
(Raksasas),] and [the Naga lords] he should spread on the two altars the skins of 
a fighting bull and a cat'. Cf. Varahamihira Brhatsamhitd 47.75-76, on the royal 
pusyasndnam: gatvd dvitiyavedim samupavisec carmandm upari raja \ deydni caiva 
carmdny upary upary evam etdni || vrsasya vrsadamsasya ruros ca prsatasya ca | 
tesdm upari simhasya vydghrasya ca tatah param; and Visnudharmottara 2.21.35 
on the brahmanical royal consecration (rdjydbhisekah): vrsasya (corr. wrkasya Ed.) 


The Saiva Age 

recitation of the Mantra text of sixteen verses prescribed for the brahmanical 
prototype when the water of consecration is poured over the king's head, 597 and, 
after the ceremony is complete, the king's return to his palace in full military pa- 
rade, mounted on an elephant or white horse, preceded by the royal banners, and 
showered with parched rice by the women standing on the roofs of the mansions 
along his route. 598 

Just as this brahmanical rite is subsumed within the Saiva process of initia- 
tion and consecration, so its outcome, the king's entitlement to rule as guardian 
of the brahmanical social order now entails the additional requirement or, one 
might say, compensation to the Saivas for this descent into the mundane, that 
he should ensure that the authority of brahmanical prescription be subsumed 
within, and subordinate to, that of the Saiva scriptures, an injunction supported 
by the promise that by enforcing this hierarchical relationship he will secure the 
stability of his rule and kingdom, implying that by neglecting to do so he will 
bring about their collapse. 599 

vrsadamsasya dvipinas ca bhrguttama | tesam upari simhasya vyaghrasya ca tatah 

597 Naimittikakarmanusamdhana ff. 78rl-79rl (interrupted by the loss of a folio), 
beginning (4.168-169): loke vede prasiddha<m>s ca vipran etarhi pathayet | 
abhisekasisah (corr. : abhisekas ikhah Cod.) slokan rsiprokta<m>s ca tad yatha || 
suras tvam abhisincantu ye ca siddha<h> puratanah \ brahma visnus ca sambhus 
ca sakradyas ca marudgandh || . . . . These verses are prescribed for this purpose by 
Varahamihira in the first half of the sixth century in Brhatsamhita 47.55-70. 

598 Naimittikakarmanusamdhana f. 84r2-5: arudho bhadramatahgam aihava 
vajinam sitam || atapatrena subhrena hemadandena *cdrund (conj. :ca + + 
Cod.) | *nigrhltatapah (conj. : + + hltatapah Cod.) svetair vljyamanas ca 
*camaraih (em. : caparaih Cod.) || caturahgabalopetah puratah ketumalaya 
|| astavighno 'nukulena dhutaya + + *vayuna (diagn. conj.:+ + + Cod.) | 
saudhagravedikasthabhih kulapatnlbhir adarat || prayuktam lajavarsam ca 
manyamano *bahupriyam (conj. wahapriyam Cod.) | praviset svapuram *pauraih 
(conj. :pau + Cod.) + + + + vikasibhih. 

599 Mohacudottara f. 21v— 22r (4.276—281): srutismrtipuranani agama dharmadesakah 
| etair yo vartate raja, sa rajyam bhunjate ciram || 277 puranam badhyate vedair 
agamais ca taduktayah | samanyam ca visesam ca saivam vaisesikam vacah || 
278 badhyabadhakabhavena no vikalpyam vicaksanaih \ yad yathavasthitam vastu 
sarvajhas tat tad avadet || 279 agamanam bahutve tu yatra vakyadvayam bhavet 
| kim pramanam tada grahyam pramanam sahkaram vacah || 280 *granthad 
granthantaram tika (?) sapeksanirapeksayoh | samadhanam tayoh karyam 
arthapattyadisadhanaih || 281 evam jnatva suradhyaksa nirvrtim paramam vraja \ 
evam dharmanvite rajhi svarastre sarvada sivam '[The sources] that teach religious 
duty are the Vedas, the Dharmasastras, the Puranas, and the Agamas. The Puranas 
are outweighed by the Vedas and the teachings of the latter by the Agamas. The 
common and the special, the latter being the teachings of Siva, are related so that 
the second outweighs the first. The learned should have no doubt about this. [For 
it is] all-knowing [Siva that] has taught everything as it truly is. When, there be- 
ing a plurality of scriptural authorities, there are two [contradictory] text-passages 


Genesis and Development of Tantrism 

The Saivas also adapted the theory of their ritual practice to enable them to 
claim that those rulers who underwent their ceremonies would be empowered in 
their efforts to maintain their supremacy and extend it through conquest. The 
ceremony of initiation had been conceived as the means of obtaining liberation 
and was always presented in these terms in theoretical texts. But a fifteenth- 
century Kashmirian scholar can proclaim in a eulogy of his patriline that by 
receiving initiation from one of his ancestors kings had expelled their enemies 
and long enjoyed distinguished reigns. 600 Similarly, an inscription of the twelfth 
or thirteenth century from Hariyana tells us that the effect of the initiation of 
King Surapala was to give him power beyond that of all his rivals. 601 It adds 
that if his Guru Murtigana initiated a brahmin, a king, or his minister he thereby 
made them [respectively] the repository of knowledge, the master of all the earth, 
and the foremost of men. 602 In the Malkapuram inscription of A.D. 1261 we 
are told that the effect of the initiation given by Visvesvarasiva to the Kakatiya 
prince Rudradeva was to make the might of his [right] arm, that is to say his 
valour in battle, shine more brilliantly 603 The same notion is apparent in the 
great Mebon inscription of A.D. 953 of the Khmer monarch Rajendravarman. 

[one non-Saiva and the other Saiva] and the question of which is valid arises one 
must privilege the teaching of Siva. The two should be reconciled, as respectively 
dependent and independent [in their validity], by means of implication and other 
exegetical tools, *[on the evidence of] the texts [themselves in which those state- 
ments occur], related texts, and commentary (?). Having understood this, Indra, 
achieve the highest bliss. Provided that the king adheres to religion in this manner, 
his kingdom will always prosper'. 

600 Rajanaka Sitikantha, Rajanakavamsaprasamsa, v. 5ab: tasmdd yodhagurur 
babhuva bhagavan samprapya dlksam yatah \prajyam rajyam apastavairinikaras 
cakrus dram bhubhujah 'His son was the Venerable Yodha. When kings received 
initiation from him they drove off all their enemies and had long and outstanding 
reigns'. For the probable identity of these kings see Sanderson 2007a, p. 397. 

601 EI I, pp. 61-66, 11.12-13.: tadbhaktimdn murtigano gunindro (corr. :gunimdro 
Ep.) babhuva bhupalahrdabjasuryah | saddlksaya yasya sa silrapaladevo 
babhuvapratimaprabhavah 'Then there was his devotee Murtigana, foremost of the 
virtuous, the sun that opened the lotus that is the heart of the king, by whose excel- 
lent initiation Surapaladeva became [a king] whose might was unequalled'. 

602 Ibid. 11. 13-14 (continuous with the passage cited in the preceding note): . . . vi- 
pram bhumipatim tadlyam athavamatyam sayam diksayet | tarn tarn bodhanidhim 
samastaprthivinatham pradhanam nrnam sthanum patrinam atanot tarum iva 
sriyajnavalkyo munih 'Any brahmin, king, or minister that he initiated he made 
the repository of [all] knowledge, lord of the whole earth, and the foremost of men, 
just as the sage Yajnavalkya caused a tree, a [mere] plant, to burst into leaf. When 
the dissolute king Supriya contemptuously refused the sacred water and grain that 
Yajnavalkya had brought to the palace to restore his health, Yajnavalkya sprinkled 
them on to a rotten tree and departed. Seeing that the dead tree immediately burst 
into leaf the king tried without success to have him return. 

603 Pantulu 1930, v. 22: srivisvesvaradesikendrasivahastodbhasidorvikramas. 


The Saiva Age 

In a passage describing his marching forth to war it speaks of the ceremony of 
[Saiva] Mandala initiation as intensifying his brilliance, a statement that in the 
context must be taken to refer to his power to conquer his enemies. 604 

Nor was it only the theory that was adjusted to suit their patrons. According 
to the Brhatkalottara the Saiva Guru was to close the initiation ceremony by 
giving abhisekah to the horses, elephants, chariots, and soldiers of the army 
by sprinkling them with the water from the vase of the Weapon-Mantra {astra- 
kalasah), one of the two main vases prepared in the course of the ceremony, "in 
order to remove all obstacles and to ensure victory in battle". 605 The Saivas also 
created a double of their ritual of post-initiatory consecration {abhisekah) to be 
performed for the king before he entered the fray 606 A much elaborated form of 
this 'consecration for victory' {jay abhisekah), involving Sakta Saiva rather than 
Saiva Mantra-deities and one thousand vases, is taught in the 248 verses of the 
27th chapter of the Uttarabhaga of the Lihgapurana. 

They also offered a wealth of apotropaic, invigorative, and hostile Mantra- 
rites that could be performed on demand for the benefit of the realm, to promote 
the success of royal patrons, and to frustrate their enemies. The evidence for such 

604 The Mebon inscription (in FlNOT 1925 [=K. 582], pp. 309-352), w. 39-40: itas ta- 
to vidyud ivddyutac chris tdvan nrpdndm pracald prakrtyd | ramyd sarat prddur 
abhun na ydvad yadiyaydtrdsamayo nirabhrd || 40 tivrdstranirdjanardjitasrir di- 
pto mahdmandaladiksayd yah | vidydhgamantrais ca krtdtmaguptih asa[dhaya]t 
siddhim uddrabhutim 'The fortune of kings, [though] unstable by nature, did not 
flicker here and there like lightning until the charming, cloudless autumn appeared, 
the season of his marching forth. His splendour enhanced by the lustration of his 
mighty weapons, he himself [made more] brilliant by initiation before the Great 
Mandala [of Siva], his person protected by the Vidyaftga Mantras, he accomplished 
the Siddhi of total success.' 

605 Brhatkalottara A, f. 45v2-3 (22.24c-25b): hastyasvaratha*yodhdndm (em. :yo- 
dhydnd Cod.) secanam astravdrind \ kartavyam vighnasamanam samgrdme jaya- 
kdranam 'He should [then] consecrate the elephants, horses, and soldiers with wa- 
ter from the Weapon[-vase] to remove obstacles and [so] bring about [the king's] 
victory [in war]'. 

606 Kirana f. 52v (27.23c-25b): prokto 'yam abhiseka<h> syd<d> vijaydrtham 
nrpasya ca | 27.24 saubhdgyajananam mukhyam grahaplddnivartakam | sarva- 
sampat*pradam srldam (corr. : pradd srldd Cod.) yasoklrtivivardhanam || 27.25 
sdntipustikarah proktah seko 'yam vighnandsakah 'This consecration that I have 
taught may also be performed to ensure a king's victory. It is the principal 
means of bringing about good fortune. It removes oppression by possessing spir- 
its. It bestows all success and wealth. It augments [the king's] fame and rep- 
utation. I have also taught it as the means of warding off ills, restoring vital- 
ity, and eliminating obstacles'; Cf. Siddhdntasdrapaddhati: evam anenaiva vi- 
dhind rdjyakdmasya bhrastardjyasya putrakdmdydh saubhdgyakdmdyd abhisekam 
kurydt 'Following this same procedure he may perform the consecration for one who 
desires sovereignty, for one who has lost his kingdom, and for a woman who desires 
a son or good fortune'. 


Genesis and Development of Tantrism 

rituals in the scriptural literature of the Saivas, especially in its Sakta Saiva 
texts, is pervasive. 607 There is also historical evidence of specific performances. 
For example, an inscription of the fifth year of the reign of the Cola Rajadhiraja 
II (r. 1163-1179 or 1166-1182) from the Tiruvalisvara temple at Arppakkam 
near Kancipuram 608 tells us that when an army from Sri Lanka had invaded 
the Pandya country, plundered the treasury of the temple of Ramesvaram, and 
interrupted the cult of Siva there, the emperor, fearing that the war might spread 
approached a certain Jnanasivadeva of Gauda, who can be seen from his name 
to have been a Saiddhantika Saiva Guru, to free the country from this menace 
by ritual means. The Guru, we are told, then worshipped Siva for this purpose 
for twenty-eight days continuously, and it was reported subsequently that these 
'attackers of Siva' (sivadrohi) had indeed been defeated. The Badaun inscription 
of Lakhanapala praises the Rajaguru Murtigana for his expertise in "the great 
rites of subjection and attraction" (1. 13: vasyakrstimahavidhananipunah); and 
Hrasvanatha, a Kashmirian Guru of the Kalikula who also held office as the 
minister of peace and war under Yasaskara (r. 939-948), performed a ritual to 
kill his king and other rituals to cause dissension and immobilize, presumably 
directed against an invading army 609 

Just as the Guru imbued the king through the ceremonies of initiation 
and consecration with the numinous power of Sivahood in the exercise of his 
sovereignty, so the Saiva rites by which the Guru assumed his office ensured 
that he, as Siva's agent among men, was imbued with the numen of royalty. As 
in the brahmanical consecration of a king, in which the royal astrologer was to 
provide him with the royal elephant, horse, throne, parasol, fly-whisk, sword, 
bow, and jewels, 610 so at the time of a Guru's consecration he received from his 
predecessor the non-martial symbols of sovereignty (rajangani, rdjacihndni), 
such as the turban, crown, parasol, sandals, fly-whisk, elephant, horse, and 
palanquin. 611 To these we may add the throne supported by sculpted lions 

607 For some examples see Sanderson 2007a, p. 281, fn. 166. 

608 ARE 20 of 1899, SII 4:456; ARE 1899, §§23-38 (partial translation in §34). 

609 See Sanderson 2007a, pp. 280-291; 2007b, pp. 295-296. 

610 Visnudharmottara 2.4.18c-20b: tato 'bhisekasambharams tasya kuryat sa daivavit 
| kunjaram turagam kuryat tasya rajfiah pariksitau | bhadrasanam ca chattram ca 
valavyajanam eva ca \ khadgaratnam taiha capam ratnani vividhani ca. 

611 Bhojadeva, Siddhantasarapaddhati f. 41v (< Svacchandatantra 4.470): usnisa- 
makutacchatrapadukacdmarahastyasvasibikadirdjahgani . . . dattva. Svacchan- 
datantra 4.70b has a throne or seat (chatram padukam dsanam) where Bho- 
jadeva has a fiy-whisk, but his account agrees with that of the Svacchandatantra 
as transmitted in Nepalese and Grantha manuscripts. Thus NAK MS 1-224, 
f. 48r3: usnisamakutadyams ca cchatrapadukacdmarah | hastyasvasibikadyams 
ca rajangani asesatah; and IFI T. 1032, p. 96: usnisamakutadyams ca 
chatracdmarapadukah \ hastyasvasibikadyams ca rajangani asesatah. In 


The Saiva Age 

(simhasanam) so intimately associated with kingship in the Indian tradition. 612 
For a manual for royal initiation, the AmrtesadTksavidhi, instructs the king 
to reward his Guru with gifts that should include golden jewellery set with 
rubies and pearls, a pair of jewelled sandals, a parasol, two white chowries, 
an elephant, and also a golden lion-throne; 613 and the Malkapuram inscription 
of A.D. 1261 describes Visvesvarasivacarya sitting on such a throne by virtue 
of his office as the Saiva Guru of the Kakatiya king Ganapati of Warangal (r. 
1199-1261), 614 decked out in royal splendour, "with his mass of tawny locks 
adorned with a diadem trembling [as he speaks], with the full-blown lotus of 
his face radiating blessings, with his pearl ear-rings striking the tops of his 
shoulders [as he moves his head from side to side], entrancing with his strings 
of pearls". 615 

Furthermore, according to the prescriptions of the Saiva scriptures the 
residence to be built for the Guru by his royal disciple was in many respects 
similar in its layout to the royal palace. It included, for example, an arsenal 
for the storage of weapons of war. 616 That Gurus should have needed the 

Lihgapurdna, Uttarabhdga, 27.259—261 the attributes of kings (nrpacihnani) are 
"the conch, the fly-whisk, the drum etc., a moon-white parasol, a palanquin, and the 
war-banner" (sahkhacdmarabheryddyam chattram candrasamaprabham | sibikdm 
vaijayantim ca sddhayen nrpateh subhdm | rdjydbhisekayuktdya ksatriydyesvardya 
vd | nrpacihnani ndnyesdm ksatriydndm vidhiyate). 

612 For an image of such a throne see, e.g., the eighth-century metal Tara from Sirpur 
(Sripura) in HUNTINGTON 1985, plate 30. The notion that the throne is the very 
embodiment of sovereignty and imparts its power to the enthroned is already found 
in the Vedic literature, in the Satapathabrdhmana ( (Gonda 1966: 45-46): 
dsandydm abhisincati | dsandl sad vai sdmrdjyam sdmrdjyenaivainam sdmrdjyam 
gamayati 'He consecrates him by affusion on the throne. The throne is indeed true 
sovereignty. Through [this] sovereignty he causes him to achieve sovereignty'. 

613 Amrtesadiksdvidhi f. 16v2— 3: 37 pascdd gurur daksaniyah svarnabhdraih *su- 
vistaraih (em. : suvistaraih Cod.) | mdnikyamuktdkhacitair alahkdrais ca 
adbhutaih | 38 navaratnamayair ddntais tathd vai ratnapdduke | haimam 
simhasanam chattram dattvd vai cdmare subhe | 39 manimuktdsvandgendra-ustra- 
mesagavddibhih | ksetragrdmddivisayair mandalais ca subhair varaih 'After that 
the Guru should be rewarded with extremely large quantities of gold, with mar- 
vellous jewellery set with rubies and pearls, made of the nine jewels, and of ivory, 
and, having given him a pair of jewelled sandals, a golden lion-throne, two white 
chowries, with jewels, pearls, horse, elephants, camels, rams, cows and the like, 
fields, villages and the like, districts, and fine provinces.' 

614 Pantulu 1930, v. 38d: tasmin ganapatyadhisagurutdsimhdsanddhydsini srivisve- 
svaradesike 'While the Guru Visvesvara[siva], occupies the lion-throne of his office 
as Guru of King Ganapati'. Note also the reading chatrapddukam dsanam 'parasol, 
sandals, and throne' in the Kashmirian text of Svacchandatantra 4.470. 

615 Pantulu 1930, v. 39: tvahgatpihgajatdkiritam udayasmerdravinddnanam muktd- 
kundalatdditdmsasikharam hdrair manohdrinam \ vidydmandapavartinam gana- 
patiksmdpdladiksdgurum srivisvesvarasambhum Tksitavatdm te caksusi caksusi. 

616 Mayasamgraha 5.182ab: dhanuhkhadgasarddini vidadhydt tu grhaksate; 


Genesis and Development of Tantrism 

means of warfare may surprise. But a fragmentary inscription of the late tenth 
century from Kadwaha in the Guna District of Madhya Pradesh relates that 
when hostile forces had invaded the region and the king had been slain, the 
Saiva ascetic Dharmasiva, abbot of the Aranipadra monastery, went into battle 
and routed the enemy through his skill as an archer, at the cost of his own 
life. 617 Nor is this an isolated instance. From the Jubbulpore stone inscription 
of Vimalasiva, Rajaguru of the Kalacuri kings Jayasimha (r. c. 1163-1188) and 
Vijayasimha (r. c. 1188-1210), we learn that the activities of his predecessor 
Kirtisiva, Rajaguru of Narasimha (r. 1153-1163), extended beyond the spiritual 
to those of a military commander who expanded his monarch's realm and added 
to his own through the appropriation of temples in the territories gained. 618 

Pihgaldmata f. 71rl-2 (10.28c-31): grhaksate grham caiva sastrasamsthdpandya 
tu | khadgabdnadhanus caiva kuthdro mudgaras tatha | cchurikd kuntadamtas ca 
citradandas tathaiva ca | lakutam sakti pasas ca kanayah sulapatrakah | cakrdsi 
gadavajras ca ahkusas ca kupattisah | evamddydni cdstrdni phardni vividhdni ca \ 
sthdpitavydni devese grhe grhaksatasya tu. The term grhaksatah here denotes [the 
deity of] a segment immediately to the east of its centre of the southern edge of the 
square plan. In the last verse I take phara- to be a variant of sphara- 'shield' from 
Iranian (Old Persian spara-barai 'shield-bearer'; Persian ispar 'shield'). 

617 EI 37:20, 11. 10-16. The inscription is fragmentary, but this much of its meaning is 
clear: while the ascetic Dharmasiva was in the monastery at Aranipada (elsewhere 
called Aranipadra) performing austerities (tendranipadam nama krtampadam ani- 
nditam . . . dattvdranipade . . . tasya dharmasiva ity abhavaj jitatma sisyah . . . tasya- 
srame vardhayatas tapamsi [11. 10-12]) a ruler called Gobhata came there with a 
force of elephants (tatrajagamonmadasindhuranam balena bhupah kila gobhata- 
khyah [1. 12]). Someone, perhaps the local ruler, was killed by this king ([nrjpena 
paragatasuh sahasa papata [11. 12-13]); and he, evidently Dharmasiva, wept with 
compassion for a while when he heard the news (tasyavagamya sa katham karuna- 
vimuktabaspah ksanam [1. 13]), then, flying into a rage (tad anu kopavipdtaflaksah] 
[1. 13]), went into battle, a veritable Siva on earth, armed with a bow *that had come 
[down to him] from Prabhava[siva?] (?) (atha prabhavagatakarmukena banais ca 
diptah sa dharavrsdhkah [1. 14]), and, like Siva in his Tripurantaka embodiment, 
routed the whole army of the enemy before ascending to the incomparable world 
[above] in a shower of flowers scattered by Indra's celestial nymphs (dtta[sva]lilas 
tripurdntakasya . . . sakalam api sa jitvd sdtravam sarvakalpah | surapatiramanl- 
ndm puspavrstydvakirnah puram anupamfam] ... [1. 15]). The poet refers here to 
the reward conventionally attributed to a warrior who dies bravely when fighting to 
protect his country; see, e.g., Mahdbhdrata 8, supplementary passage 14, 11. 31-34; 
13, supplementary passage 15, 11.1358-1361. 

618 EI 25:33 (A.D. 1174), vv. 23—24: na syandanam vasumati na ca candrasuryau cakre 
na sdrathir abhut sa ca visyayonih \ nesur harih parapurdni tathdpi bhasma cakre 
yatah sa iti kirtisivah sphutam sah \\ yasobhir induvisadais tathaivdrivikarsitaih | 
apupurat sa sarvdsd vivekakusumair iva 'He was manifestly [worthy of the name] 
Kirtisiva [Temple/Fame-Siva]. For he [was a Siva in as much as he] reduced the 
cities of his enemies to ashes [just as Siva did to the cities of the three demons] even 
though his war chariot [unlike Siva's] was not the earth, the sun and moon were 
not its two wheels, its driver was not Brahma, and his arrow was not Visnu; and he 
filled all the directions with the moon-white temples that he had wrested from his 


The Saiva Age 

Kings rewarded their Gurus with the donation or construction of monas- 
teries (mathah) and with grants of revenue from designated lands with which 
they themselves constructed and endowed such institutions. Thus in the first 
half of the ninth century the Rajaguru Purandara founded two monasteries 
in Gwalior, one at Mattamayura and a second at Aranipadra, using the funds 
he had received from king Avantivarman as the daksina for performing the 
king's Saiva initiation, for which purpose he had been persuaded to move to 
Mattamayura, probably from Malava. The wealth received is described in 
the inscription that records these events as "[the revenue of] the most valu- 
able portion of his kingdom". 619 Similarly, when the Kalacuri Yuvarajadeva 

foes, just as he did with the [white] blossoms of his Vivekas'. My translation finds 
a reference to [lost] works by Kirtisiva entitled Viveka, presumably commentaries 
on Saiva texts. It is possible that the poet refers not to works but to Klrtisiva's 
spiritual insights (vivekah). 
619 Ranod inscription, EI 1:41, vv. 10—15: tasmat purandaragurur guruvad garimnah 
prajnatirekajanitasya babhuva bhumih | yasyddhunapi vibudhair itikrtyasamsi 
vyahanyate na vacanam nayamdrgavidbhih || 11 vandyah ko 'pi cakdsty acintya- 
mahimd tulyam munir bhdsvatd rdjann uttamasabdapurvasikhardbhyarnam 
prakirnadyutih \ diksarthiti vaco nisamya sukrti cdroktam urvipatir yasyehdna- 
yandya yatnam akaroc chriman avantih purd || 12 gatvd tapasyantam upendrapiirve 
pure tadd srimadavantivarma | bhrsam samdrddhya tarn atmabhumim kathamcid 
anlya cakaraputdm || 13 athopasadydpya ca samyag aisim dlksdm sa dakso guruda- 
ksinartham | nivedya yasmai nijarajyasaram svajanmasdphalyam avdpa bhupah || 
14 sa karayam asa samrddhibhajam munir matham sanmuniratnabhumim \prasi- 
ddham dvaridhi merukalpam srimatpure mattamayuranamni || 15 punar dvitlyam 
svayam advitlyo gunair munlndro 'ranipadrasamjnam | tapovanam sresthamatham 
vidhaya presthah pratistham paramam ninaya 'Then came the Guru Purandara, 
who as befitted a Guru had the gravity that comes from the highest wisdom, whose 
teachings concerning the duties [of Saiva initiates] have still not been surpassed 
by scholars learned in the way of discipline, whom the glorious and virtuous king 
Avanti[varman] made efforts to bring to this land because he desired to receive 
[Saiva] initiation and had heard from one of his agents that there was a certain 
holy ascetic in the vicinity of Uttamasikhara shining in unimaginable glory, shed- 
ding his radiance like the sun. Avantivarman then went to [Purandara], who was 
practising austerities in Upendrapura, and having striven to win his favour suc- 
ceeded in bringing him back to sanctify his kingdom. Then, having served him with 
devotion he duly received Saiva initiation [from him]. The wise king then presented 
him with the best part of the wealth of his kingdom as Guru's fee and so brought 
his human birth to fulfilment. In the splendid town of Mattamayura the sage then 
caused a richly endowed Meru-like monastery to be built, a treasury of jewel-like 
ascetics, the fame of which has reached [throughout the continent] to the oceans. 
This foremost of sages, himself unmatched in his virtues, built and richly endowed 
a second and most splendid monastery, [this] hermitage of Aranipadra'. I say that 
Purandara probably came from Malava because we are told here that before he was 
brought to Mattamayura he was in Upendrapura and a grant of 1110 issued by 
the Paramara king Naravarman (EI 20:11) refers to the gifting of land in a village 
in the district of Upendrapura (1. 5: upendrapuramandale), which must have been 
within his kingdom, that is to say, in Malava. It is probable that this town and 


Genesis and Development of Tantrism 

I alias Keyuravarsa (r. c. 915-945) induced Purandara's spiritual descendant 
Prabhavasiva (/Sadbhavasiva) to move to his kingdom in Chattisgarh, he 
founded for him at huge expense the great monastery at Golagi, 620 granting him 
by royal charter numerous villages and a whole well-populated town, which, 
since it is not named, was probably Golagi itself, 621 or, according to the account 

district bore the name of Upendra, the first of the Paramara kings according to the 
genealogy given by the poet Padmagupta in 11.76 of his Navasahasahkacarita. 

620 In all secondary sources, including SANDERSON 2007a (p. 274), the name of this 
monastery (mathah) appears as Golaki-. That spelling is well attested, but only in 
manuscripts and inscriptions from the Dravidian South, where the scribes, speak- 
ing languages in which voiced and unvoiced consonants are not distinguished, are 
liable to substitute k for g. We also find kolaki there. I now correct to Golagi- be- 
cause this is what I find in the earliest testimony, which comes from regions whose 
vernacular languages do distinguish these consonants, namely Nepalese palm-leaf 
manuscripts of the Kriyakdndakramavali and the Bangarh Prasasti of the time 
of Nayapala (r.c. 1027-1043) (SlRCAR 1983b, v. 6: golagyas sa mahamathah). The 
name appears as Golaggi in the Chandrehe inscription (caranaputagolaggikah). I 
identify Golagi with modern Gurgi (244° 31' N, 81° 27' E), about 12 miles due east of 
Rewa Town, in the north of the Kalacuri kingdom. This is the site of once vast Saiva 
ruins (Cunningham 1885, pp. 149-154; Meyer et al. 1908-1931, vol. 21, pp. 282- 
283; Banerji 1931, pp. 41-45). A full account of my reasons for proposing this 
location and for rejecting as groundless the widespread view that the monastery 
was in the south of the kingdom at Bheraghat on the Narmada river, close to the 
Kalacuri capital Tripuri, must be set out elsewhere. 

621 Chandrehe inscription, CII 4i:44, v. 5: tato madhumatlpateh krtamahatapah- 
samcayah prabhavasiva ity abhut sakalasaivacudamanih | anekanrpavanditah sa 
yuvarajadevena yas tapodhanapatih krtas caranaputa*golaggikah (my reading : 
golagnikah MlRASHl, Banerji [EI 21:23]) 'Then after the abbot of Madhumati 
came that crest-jewel of all the Saivas called Prabhavasiva, who had accumulated 
vast power through his asceticism and was revered by many kings. He purified 
Golaggi [=GolagI] with his feet after being appointed by Yuvarajadeva as overlord 
of the ascetics [of the monastery at that placel'; and the Gurgi inscription, EI 
22:21, vv. 6—7: tasyakhilaksitipatipranatottamahgacudamanidyuticayarcitapada- 
pithah | sisyo babhuva bhuvanatrayakirtaniyah srimatprabhavasivanamamunir 
manlsi || anlya yam sahajavasanaya nayajhah srimugdhatuhgatanayo yuvaraja- 
deva}} | sattvopakarabhavaduttamakirtihetor agrahayan matham anantadhana- 
pratistham 'His disciple was the glorious and learned ascetic Prabhavasiva, wor- 
thy of celebration throughout the three worlds, the pedestal beneath whose feet 
was honoured by the dense rays of the crest-jewels on the heads of all the kings who 
prostrated themselves before him. Yuvarajadeva, the son of Mugdhatunga, skilled 
in policy, brought [him to his kingdom prompted] by an inborn predisposition and 
had him accept a monastery that he established [for him] with infinite wealth'. The 
damaged vv. 35-40 at the end of this inscription list the places that the king made 
over to Prabhavasiva: [sthanam] . - . . . - . ya kirtanifyam] punyanvitaya mu- 
naye svayam arcitaya | - - nam ullikhitafsasana - . - - keyu]avarsanrpatih [svayam 
ajahara] \\ 36pakk + + . - - [tarn?] tatha sarasadollakam | vakkadollakarajyauddhe 
ko + +[na]sapundika || 37 + + + + . — \- + +puram khatollika \ . nakalabhirapalli + + 
+ + sarasvatl || 38 [etesam] dvadasakah ca kavacaksetram eva ca | sdmantapatakas 
caiva vata + + . - . + || 39 + + + yaftallapatT] sasanam [sajtram ity api | sa + + bhad- 
dhaci[u]ra [kusu?]mva ca ku[kku]diya || 40 rajogrdmanvitafn etan sajsanatvena dat- 


The Saiva Age 

of the Malkapuram inscription, gave him a vast reward which that ascetic, 
after he had himself founded the monastery, transferred to it as its endow- 
ment. 622 In the next generation the Kalacuri Laksmanaraja II (r. c. 945-970) 
brought in Hrdayasiva and gave him the monasteries attached to the temples of 
Vaidyanatha and Nohalesvara, the second of which Hrdayasiva passed on to his 
disciple Aghorasiva; 623 and the Bangarh Prasasti reports, as we have seen, that 

tavdfn] | + + + + . [siddhdjntapdragdya gariyase || puram paurajandkirnam + + + + 
samastakam | bhaktyd samarpaydm dsa sdsanatvefna bhuipatih. 

622 Pantulu 1930, vv. 25c-26: tasmai nihsprhacetase galacuriksmdpdlacuddmanir 
grdmdndm yuvardjadevanrpatir bhiksam trilaksim dadau || 26 krtvd 
sa saivamunir adbhutasilamurtih srigolakimatham uddram uddttacittah \ 
[tajsydkarasya nrpadesikamauktikdndm vrttim cakdra sakaldm api tarn trilaksim 
'To that [ascetic] whose mind was free of all craving the king Yuvarajadeva, that 
crest-jewel among the Kalacuri monarchs, gave a 300,000 endowment of villages. 
That Saiva ascetic, the noble-minded embodiment of extraordinary good conduct, 
built the great Golaki [Golagi] monastery [there] and then made over the whole of 
that 300,000 living to that [monastery, which, ocean-like, has become] the source 
of [many] pearls in the form of Rajagurus'. MlRASHl (CII 4i, p. clviii) interprets 
the words grdmdndm bhiksam trilaksim 'a 300,000 endowment of villages' to 
mean that 300,000 villages were given to Prabhavasiva and points out that if the 
report is correct it indicates that "the king assigned to him one third of the total 
revenue of his home province of Dahala, which, according to tradition, comprised 
nine lakhs of villages". This would indeed be a vast endowment, so vast indeed 
that I find it hard to accept his interpretation. The Gurgi inscription mentions 
only about twenty villages and a town and the Malkapuram inscription need 
mean only that the endowment [consisting of the revenue capacity of these places] 
was valued at 300,000 of some unspecified monetary unit. This alternative was 
already considered by Pantulu, the first editor of the Malkapuram inscription. 
For though he proposed the interpretation later adopted by Mirashi, he saw the 
difficulty it entails (1930, p. 52): "The founder of the monastery was one Sadbhava 
Sambhu who obtained a gift of three lacks [sic] of villages (or was it a villages [sic] 
fetching an income ofNishkas (coins)?) from the Kalachuri king Yuvarajadeva and 
gave away those villages to the Matha as an endowment". In favour of this more 
realistic reading is a parallel expression seen in an inscription of the sixth century 
from a site near Mrohaung in Arakan. There we learn of the gift to a Buddhist 
monastery of a trisdhasriko grdmah {EI 37:13, 1. 13: dehguttandmd ttrisdhasriko 
grdmo nisrsto), which can only mean 'a village which has [a revenue yield of] 3000'. 
As the editor, D.C. Sircar points out (p. 63), this refers "apparently to the revenue 
income in the standard coin". 

623 Bilhari inscription, EI 1:31, vv. 56-58: 56 kim stuyate 'sau munipuhgavo 'thavd 
sricedicandro nrpatih krtddarah | sadvrttadutaprahitair updyanaih pradarsya 
bhaktim vidhindnindya yam \\ 57 srimallaksmanardjo 'pi tasmai sutapase svayam \ 
matham srivaidyandthasya bhaktiyuktah samdrpayat \\ 58 svikrtydpi munir bhuyo 
matham srinauhalesvaram \ aghorasivasisyasya sddhuvrttasya dattavdn 'Or rather 
why should I praise that foremost among ascetics? [It suffices to report that] king 
Laksmanaraja, the moon of the Cedi dynasty, brought him [to his kingdom] after 
earnestly showing his devotion to him through presents sent by virtuous envoys, 
and then out of his devotion freely bestowed on that [saint] of great austerity the 
monastery of Vaidyanatha. The ascetic also accepted the monastery of Nohalesvara 
and then gave it to his virtuous disciple Aghorasiva'. 


Genesis and Development of Tantrism 

the Pala emperor Mahipala I (r. c. 977-1027) bestowed a lofty gilded monastery 
on the Guru Indrasiva at Sivavati near Kotivarsa. 624 

Moreover, we have several records of Gurus using their resources in- 
dependently to establish further monasteries. Thus Prabhavasiva's disciple 
Prasantasiva built a monastery at Chandrehe for ascetics devoted to med- 
itation 625 and a hermitage on the banks of the Ganges at Benares. 626 His 
disciple, the Rajaguru Prabodhasiva, also built a monastery at Chandrehe; 627 

624 Bangarh inscription, SlRCAR 1983b, v. 9: srlman indrasivah sphutam hari- 
haraprayam sivendrakrtim bibhrad vamsavibhusanam samabhavac chisyo 'sya 
punyatmanah \ yasmai kancanapunjamanjuracitaprasadamerusphuratkailasabha- 
matham daddv iha mahipdlo nrpas tattvavit 'The disciple of that [Guru] devoted to 
piety was the illustrious Indrasiva, an ornament of his lineage, who did indeed have 
an appearance [matching his name, in that it was one] that embodied both Siva and 
Indra [=Upendra, i.e. Visnu] as though it were an image of Harihara [in which 
Siva is both himself and Visnu in a single body]. To him king Mahipala, [once he 
had become through initiation] a knower of [ultimate] reality, gave in this place a 
monastery that resembled Mt. Kailasa, radiant with its Meru-like towers beauti- 
fully wrought with much gold'. 

625 Chandrehe inscription, CII 4i:44, vv. 6a, 7: prasantasivacandramds tad anu tasya 
sisyo 'bhavat ... 7 sa sonanadasamgame bhramarasailamule 'tulam priyalavana- 
samkule phalamrnalakandasanah | cakdra viditam janair munisakhah prasanta- 
sramam svapadapadapahktibhih pavitabhutalo yah krti 'The successor of [Prabha- 
vasiva] was his disciple, the moon-like Prasantasiva. . . . Eating [nothing but] fruits, 
lotus stems, and bulbs, that wise friend of ascetics built the famous hermitage with 
his name [the Prasantasrama] at the foot, thick with a forest of Priyala trees, of the 
Bhramara hill, at the confluence of the river Son, purifying the earth with the lines 
of his foot-prints'; and the Gurgi inscription, EI 22:21, vv. 8 and 13: tasydmalena 
tapasa ca vivardhamanavidyabalena ca samastajagatpratitah | sisyah prakama- 
kamaniyagunaikadhama srimatprasantasivandmamunir babhuva \\ ... 13 dahotti- 
rnasuvarnadanasamitadravyarthisarthasprhah siddhasthanam acikarat tad apa- 
ram yah sonatlropari | yasmin yogajusah pravisya niyamadhvastdntarayadhayah 
santah siddhasamadhayo 'cchamatayo gacchanti mukteh padam 'The disciple of 
this [Prabhavasiva] was the ascetic Prasantasiva, who was known to all for his 
unblemished austerity and the power of his ever growing knowledge, the unique 
abode of the most desirable of qualities. . . . [13:] He, who quenched the desire of a 
multitude of people in need of funds with fire-refined gold, built another [monastery 
as] a seat of Siddhas on the bank of the river Son, where masters of Yoga enter, abol- 
ish the torment of [all] hindrances through their ascetic restraint, and, when they 
are at peace, having achieved perfect concentration, reach with pure awareness the 
goal of liberation'. 

626 Gurgi inscription, EI 22:21, v. 14: tlrthasnananisevanodyatadhiyam atyan- 
tavisrantaye yas tat karitavan munih surasarittire tapahsthdnakam | yat samsevya 
mahesvararcanaratd varanaswasino manyante bhavasagaram gurum api ksinam 
yatha [gospajdam 'That ascetic had a hermitage built on the bank of the Ganges for 
the complete repose of those whose minds were devoted to the practice of bathing 
at its Tirthas. By resorting to it those living in Benares who are devoted to the 
worship of Siva consider the ocean of transmigratory existence, vast though it is, to 
have dwindled into a mere puddle'. 

627 Chandrehe inscription, CII 4i:44, v. 16ab: gurukrtasuragarad arad amum matham 


The Saiva Age 

Patarigasiva, a spiritual descendant of Purandara through another line, 
built a monastery in Gwalior at a site now unknown; 628 and the Rajaguru 
Visvesvarasiva, after receiving a village in Andhra from the Kakatiya Queen 
Rudradevi, built a monastery there and renamed the village Visvesvaragolaki 
after both himself and the original home of his preceptorial lineage in Chat- 
tisgarh, dictating that only a Guru of this lineage, one consecrated by another 
Guru of the same (golakivamsyakrtabhisekah), should be allowed to preside over 
his foundation. 629 According to the same source he also established monasteries 
in Kalisvarapura, Mandrakutanagara (v. 82), and Isvarapura (v. 85), no doubt 
under the same conditions. 

In this way there developed a far-reaching network of interconnected seats 
of Saiddhantika Saiva learning. Figures at the summit of this clerical hierarchy 
thus came to exercize a transregional authority whose geographical extent could 
be greater than that of any contemporary king. Visvesvarasiva while holding of- 
fice as the Rajaguru of the Kakatiya Ganapati is said also to have been the Guru 
of the Kalacuri king, the Cola king, and the king of Malava; 630 and praise of Saiva 

unnatam svakam iva yasah subhrabhrabham visalam acikarat 'Near the temple 
built by his teacher he built this broad and lofty monastery that resembles a white 
cloud, as though it were his own fame'. 

628 Gwalior Museum inscription, MlRASHl 1962, v. 40: matham devakulam kupas 
tadaganam ca pahcakam | pra[ka]ro vatika ... 'A monastery, a temple, wells, five 
reservoirs, a circumvallation, *an orchard (?) 

629 Malkapuram inscription, Pantulu 1930, w. 42-45 and v. 70: 69c-72: 
devasya sattrasya mathasya tasya gramasya sarvasya ca so 'dhikari || 70 yo 
golakivamsyakrtabhisekah santah sucih saivarahasyavedi \ saivagamanam 
api paragami samtanapalah samalostahema || 71 sarvani bhutany anukam- 
pamanah krtavagahah | mahisurah silavatam purogo bhavettaram 
naisthikadesikendrah | 72 visvesvarasivacaryo dhiman rajaguruh svayam evam 
ajnapayad dhirah saiv deary asatair vrtah 'Surrounded by hundreds of Saivacaryas 
the learned and noble Visvesvasivacarya personally ordered that the superinten- 
dent of the [temple of the] god [Visvesvara], the refectory, the monastery, and the 
whole settlement [that he had established] could only be an ascetic Guru whose 
consecration [to office] had been performed by [a Guru] of the lineage of Golagi, a 
brahmin outstanding among the virtuous, tranquil, honest, one who understands 
the esoteric doctrines taught by Siva, who has mastered the Saiva scriptures, a 
guardian of his initiatory line, for whom a clod of earth and gold are of equal value, 
compassionate to all living beings, and deeply versed in all branches of learning'. 

630 Malkapuram inscription, Pantulu 1930, v. 38: sricolesvaramalavaksitipati 
rajanyacudamani yacchisyau kim atah param ganapatiksonipatir yatsutah\na 
sydt kasya made sa desikavarah saivagamambhonidhih srivisvesvaradesikah 
kalacuriksmapaladiksaguruh 'The Cola king and the king of Malava, the crest- 
jewels among rulers, were his disciples. King Ganapati too was his [spiritual] son. 
Whom does this excellent Guru not delight? The Guru Visvesvara, this ocean of 
[knowledge of] the Saiva scriptures, was the Guru that [also] initiated the Kalacuri 


Genesis and Development of Tantrism 

Gurus as venerated by a plurality of kings is common, even a commonplace. 631 

The wealth accumulated by these Gurus enabled them behave like royal 
patrons themselves, not only founding new monasteries but also bestowing land- 
grants on brahmins, rewarding poets, founding temples and new settlements, 
and providing the means of irrigation. The Badaun inscription reports that 
the Rajaguru Murtigana "honoured brahmins in abundance with many gifts 
of land that he had received due to the devotion of his royal disciple"; 632 the 
Malkapuram inscription says concerning the Rajaguru Visvesvarasiva, a na- 
tive of Gauda in eastern India: "Who can count the Gauda [brahmins] whose 
wishes he has granted, the ascetics who have received rich endowments [from 
him], the leading poets who have been delighted [with the rewards he has be- 
stowed]?"; 633 and the Bangarh Prasasti relates that Sarvasiva, the Rajaguru of 
the Pala Nayapala, gave [to brahmins] all the Great Gifts (mahdddndni) of the 
Puranic tradition, including the tulapurusadanam in which the donor gives away 
his weight in gold, an activity that increasingly became emblematic of exemplary 
kings during the second half of the first millenium. 634 His brother Murtisiva, to 

631 See, for example, in the colophonic verses of the Prdyascittasamuccaya of 
Hrdayasiva, concerning his Guru Isvarasiva (see Sanderson 2001, p. 3): asit 
tatsamtatau munih sri-lsvarasiva iti | jagatlpatibhir nrpaih pujitapddapahkajah; 
Chandrehe inscription (CII 4i:44), v. 4b, concerning Purandara: yatra puranda- 
rah krtatapd jajne gurur bhubhujdm; v. 5c, concerning Prabhavasiva: anekanrpa- 
vanditah; Bilhari inscription (CII 4i:45), v. 50b, concerning Dharmasiva: bhupd- 
lamaulimanikdntibhir arcitdnghrih; v. 51bcd, concerning Sadasiva: nrpaih \ yat- 
pddadvayam vandyam arcitam sekhardmsubhih; v. 54cd, concerning Hrdayasiva: 
nrpamukutanivistair yasya mdnikyacakrair akrta caranamulam kantam ekdntava- 
ndyam; Gurgi inscription (CII 4i:46), v. 6, concerning Prabhavasiva: tasyakhilaksiti- 
patipranatottamdhgacudamanidyuticaydrcitapddapithah | sisyo babhuva bhuvana- 
trayakirtaniyah srimatprabhavasivanamamunir manisi; and v. 17cd, concerning 
Isanasiva: srisdnasambhur akhilavanipalamaulimdldmanidyutipisahgitapddapa- 

632 Badaun inscription, EI 1:10, 1. 15: svasisyavarabhupdlabhaktilabdhena bhurind | 
bhumiddnena yo viprdn pujaydm dsa bhurind. 

633 Pantulu 1930, v. 39ab: gauddh purnamanorathdh kati kati prdptasriyas tdpasdh 
samtustdh kavipumgavdh kati kati pradhvastapdsd nrpdh. 

634 SlRCAR 1983b, v. 11. The inscription lists prthividdnam , meruddnam, visvacakra- 
ddnam, [sapta]sdgaraddnam, brahmdndaddnam, kalpavrksaddnam, [hiranyajkd- 
madhenuddnam, bhavanaddnam, grdmaddnam, goddnam, parvatdndm ddnam 
(the ten parvataddndni of the Matsyapurdna, with Meru in the centre), sakalpa- 
drumabhadraghataddnam, hiranydsva[ratha]ddnam, hiranyahasti[ratha]ddnam, 
hiranyagarbhaddnam, asvaddnam, tulapurusadanam, and srinandisvaraddnam. 
For an exhaustive presentation of the prescriptions of the Puranic and other sources 
on the "Great Gifts" see especially the fifth Adhydya of the Ddnakhanda of the 
Caturvargacintdmani of Hemadri, written while he was a minister of Mahadeva, 
the Yadava king of Devagiri (r. c. 1260-1270). The srinandisvaraddnam mentioned 
in this inscription is, I presume, the gift of a golden image of Nandikesvara that is 
to accompany the gift of a thousand cows (Caturvargacintdmani, vol. 1, p. 253). On 


The Saiva Age 

whom he handed over his office as Rajaguru is likewise praised in that inscrip- 
tion for his abundant donations to brahmins. 635 Sarvasiva's disciple Rupasiva is 
also praised there for his generosity to supplicants, 636 as is Isanasiva, the disciple 
of Prasantasiva, in the Gurgi inscription. 637 The predecessors of the Rajaguru Vi- 
malasiva receive similar praise for their pious largesse in that Guru's Jubbulpore 
inscription, and Vimalasiva himself is commended there for the support he gave 
to the brahmanical order by bestowing gifts on brahmins, and adorning the land 
with gardens, water-tanks, charitable feeding-houses (sattrani), temples, and 
houses for brahmins. 638 In the Bangarh Prasasti Vidyasiva and Dharmasiva are 

the drift away during our period from the sponsoring of Vedic (Srauta) sacrifices to 
the bestowing of the Great Gifts such as the tulapurusaddnam see DlRKS 1976. 

635 SlRCAR 1983b, v. 15cd: bhratd murtisivah sa mdnyamahimo ddnambusekair jagat 
putam yah krtavdn . . . 'His brother Murtisiva, of venerable glory, washed the world 
clean with the water he poured when making donations'. The poet refers to the rite 
of pouring water on to the hand of the brahmin recipient, or, in his absence, on to 
the ground, that must accompany any formal act of donation (Caturvargacintdmani, 
vol. 1, p. 92); and by saying that he cleansed the world with these libations he sug- 
gests that his donations to brahmins were frequent, widespread, and very numer- 

636 SlRCAR 1983b, v. 28: sisyah sarvasivasya diptatapasah sarvdrthicintdmanir . . . | 
srimdn riipasivo babhuva 'The disciple who succeeded Sarvasiva, [that Guru] of 
blazing ascetic power, was the illustrious Rupasiva, who was a wishing-granting 
jewel for all supplicants'. 

637 EI 22:21, v. 18ab: . . . \sarvdrthi\ndm yena srlr gamitopabhogapadavlm 
daurgatyaduhkhacchidd 'He caused [his] wealth to be enjoyed by all suppli- 
cants, thus ending the torment of their poverty'. 

638 EI 25:33. The inscription precedes its account of the life of Vimalasiva with some in- 
formation about the predecessors in his Guru lineage. Unfortunately the section on 
his predecessors is lacunose because of damage to the stone, with the loss or partial 
loss of some of these Gurus' names. The inscription yields the following succes- 
sion: . . . N > Vimalasiva > Astrasiva — in 11. 5-6 I read ... (1. 6) vdstrasivdbhidhanah 
where the editor, MlRASHl, reads ...(1.6) vdstusivdbhidhanah: Astrasiva is a 
Saiddhantika initiation name but *Vastusiva is not — > N? (if Astrasiva's succes- 
sor was covered in the lost v. 11) > N-siva (the first part of the name has been 
lost: . . .sivah sisyah in 1. 6) . . .N > Purusasiva, Guru of Yasahkarna (r. 1073-1123) 
> Saktisiva, Guru of Yasahkarna's successor Gayakarna (r. 1123-1153) Klrtisiva, 
Guru of Gayakarna's successor Narasimha (r. A.D. 1153-1163) > Vimalasiva, Guru 
of Narasimha's successors Jayasimha (r. 1153-1188) and, on the evidence of EI 
40:46, Vijayasimha (r. 1188-1210). Of N-siva we are told (v. 11): + sivah sisyah 
purusarthdya sampadam | gunandm ca dhanandm ca paropakrtaye param '[His] 
disciple N-siva [employed] his abundant virtues only for the accomplishment of 
the goal of human existence and his abundant wealth only for the welfare of 
others'; and of his now nameless successor we learn ... (v. 15) pritih pdtre ratis 
tirthe sthitih pathi mate satam | bhaktis bhave 'bhavat tasya samasya 'That as- 
cetic's only delight was in [giving to] worthy recipients, his only attachment was 
to holy sites, his only adherence was to the path approved by the good, and 
his only devotion was to Siva'. Of Vimalasiva we learn in v. 34cd: [yacchdjydm 
vibudhagano 'dhigamya dhatte vaidhuryam na khalu [mahotsajvodayesu 'Enter- 


Genesis and Development of Tantrism 

praised for building temples, 639 and the Rajaguru Murtisiva for building many 640 
and excavating numerous reservoirs. 641 In the Gurgi inscription Prasantasiva is 
said to have added a lofty temple of Siva at Golagi to the north of one that had 
been established there by king Yuvarajadeva; 642 and in the Chandrehe inscrip- 
tion his successor Prabodhasiva is said to have provided that place not only with 
a monastery but also with a water reservoir and a well. 643 The Gwalior Mu- 

ing the shade [provided by the parasol] of this [patron] a multitude of brahmins 
was freed from the distress [of penury] on the splendid occasions of major festi- 
vals'; in v. 38: yasydrthidvijardjadarsanavasdd ddndmbufbhir vardhate] sraddha 
[rdtridivam] varena vidhind dharmasya *tantrlr (?) iva | yo darsesv api sadaram 
dvijapatin aksinasobhdbhardn dakso yojayate suvarnavikasatsadrohinindm sataih 
'At the sight of great brahmin supplicants his faith grows day and night along with 
the [frequency with which he does] the pouring of the water of donation, in accor- 
dance with the best procedure, like a * ... (?) of religious duty. And on the days 
of the new moon [this] learned [Guru] bestows with devotion on the leading brah- 
mins, their rich adornments never diminished, hundreds of fine ruddy cows shining 
with gold [adorning their horns]'; v. 41bc: [dattam] na yan ndsti tat | pdtram tan na 
yad arcitam 'there is no gift that he did not give, no worthy recipient whom he did 
not honour'; and v. 43: udydnasarasi[sattra]prdsddadvijavesmabhih | bhumih parib- 
havaty asya na kair bhusdbharair divam 'With what rich adornments [created by 
him], with gardens, reservoirs, charitable feeding-houses, temples, and houses for 
brahmins, did [this] land not surpass heaven?' 

639 SlRCAR 1983b, v. 8ab: sisyo dharmasivas taponidhir abhut tasya vyadhdd yo 
'dbhutam prasadam bhagavattrilocanaguror vdrdnasibhusanam 'His disciple, the 
ascetic Dharmasiva, built a marvellous temple of the blessed three-eyed teacher 
[of the world] that beautified Benares'; Sircar 1983b, v. 7cd: srividydsiva 
ity asimacaritas satkirtisdkhdsataprdgbhdrasthagitdmbaro munir abhut tasmdd 
yathdrthdnvayah 'After him came Vidyasiva, an ascetic of boundless virtuous con- 
duct, in whom the lineage fulfilled its purpose, who concealed the sky with the mass 
of the countless branches of his fine temples'; vv. 16-19. 

640 SlRCAR 1983b, v. 19: mahiyasiyam na tathd mahi yathd tapasvinas tasya mahdn 
ihdsayah | tathd hi bhumih kila kirtibhir bhrtd gato na tasydsaya esa vismayah 
'This land though vast was not large enough for the ambition of this ascetic. The 
wonder is that it did not cease even when the earth was filled to capacity by his 

641 SlRCAR 1983b, v. 17ab: ...nirmitds citram diksu vidiksu yena prthivihdra*sriyo 
(conj. sriyd Ep.) dirghikdh 'Wondrously he created reservoirs in all directions as 
a beautiful garland to adorn the land'. 

642 Gurgi inscription, EI 22:21, v. 11: yena sriyuvardjakdritalasatkaildsasrhgopama- 
prdsddottaratah sumerusikharaspardhi prasiddhafm bhujvi \ sadma sthdpitam 
isvarasya *sakalatrailokyavismdpakam (trailokya corr. MlRASHl : traildkya Ep.) 
yat svargam vrajatas tadlyayasasah sopdnamdrgdyate 'To the north of the temple 
built by Yuvaraja that resembled the shining peak of Mt. Meru he built his famous 
temple of Siva. That [too] rivals the peak of Meru, causing wonder throughout the 
three worlds, a flight of steps, as it were, for his fame as it ascends to heaven'. The 
repetition of the comparison with Mt. Kailasa seems lame, but its probable point is 
that the Guru's temple was no less impressive than the king's. 

643 Chandrehe inscription, CII 4i:44, v. 16cd: anugiram atho sindhuprakhyam taddgam 
acikhanat pracurasalilam kupam cdtra prabodhasivah sami 'Then here [after 
building the monastery] the ascetic Prabodhasiva excavated an ocean-like reservoir 


The Saiva Age 

seum inscription records that Patarigasiva built a great temple of Siva 644 and 
excavated four huge reservoirs. 645 The Jubbulpore inscription records that the 
Rajaguru Vimalasiva built a temple of Siva Kirtisvara in honour of his preceptor 
and predecessor, the Rajaguru Kirtisiva. 646 A Kannada inscription recording the 
death in 931 of the Saiva Guru Tribhuvanakartaradeva alias Kaliyugarudra tells 
us that during the forty years of his rule as the pontiff of Avani in Nolambavadi 
he built fifty temples and two large water reservoirs; 647 and the Malkapuram in- 
scription records that the Rajaguru Visvesvara founded temples to house Sivas 
bearing his own name in Visvesvaragolaki, Mandrakutanagara, Candravallina- 
gara, Visvesvaranagara, Kommurgrama, and Uttarasomasila, and also that he 
founded a town with his own name (Visvesvarapura) at Ananda. 648 

The exalted status and king-like behaviour of these Gurus is reflected in the 
fact that we have inscriptions in which they have been given royal, even impe- 
rial titles. This is so with Vamadeva, also called Vamasambhu, the Rajaguru of 
a Kalacuri of Tripuri who was probably Garigeyadeva (r. c. 1015-1041), on whom 
that king is said to have transferred his status as the monarch {nijarajalaksmi) 
as payment for his service as his Guru (gurudaksina) when he set out on a cam- 

near the [Bhramara] hill and a well with abundant water'. 

644 MlRASHl 1962, v. 29: tenedam haramandiram susikharam yat sarvatah sundaram 
bhaktyd kdritam indudhamadhavalam kaildsdsailopamam | dkalpam sthiram 
astu tad bhuvi satdm dnandadam darsandd asyaivdmalam dgamat parinatim 
prdsddamurtyd yasah 'Out of devotion he had this temple of Siva built with its 
fine towers, altogether beautiful, white as the light of the moon, resembling Mt. 
Kailasa. May it endure on earth to the end of the aeon, delighting the virtuous 
when they see it. His spotless fame has been transformed to take material form as 
[this] temple'. 

645 MlRASHl 1962, vv. (30— )38: sutatam catustayam idam ruciram cirabhusanam 
mahlvadhvdh | vikatatarataddgdndm acikarac chrlpatahgesah 'Patahgasiva made 
these four lovely and immense reservoirs with beautiful banks as an enduring or- 
nament for the woman that is the earth'. 

646 EI 25:33, vv. 45-46: [a]cikarac candramauler mandiram ddardt | guror 
kirtisivasyaitat kirtaye sukrtdya ca || devdya kirtisvarasamjnitdya prdddd amusmai 
jayasimhadevah | bibhrad bhave bhaktibharam gurau ca grdmdn raveh parvani n 
+ ddydn 'He built out of reverence this temple of Siva for the fame and religious 
merit of the Guru Kirtisiva. The god [installed in it] was named Kirtisvara; and 
King Jayasimha, having great devotion both to Siva and [his] Guru, gave it [three] 
villages as * ... (?) gifts on the sacred day of the sun['s eclipse]'. 

647 EC 10, Mb:65: svasti srimad-dvanyada sthdnamam ndlvattu-varsaman did ayva- 
ttu-degulam mddi piriyav-eradu-kereya katti saka-varsam entu-nur-embatta-mur 
ddand utkrdnti geydu srltribhuvanakarttara-devam kali-yuga-rudrdhka rudra-lo- 
ka-prdptan ddam 'Hail! After governing the sacred domain of Avani for forty years 
and building fifty temples and two large reservoirs, in the Saka year 853 [the soul 
of] Tribhuvanakartaradeva alias Kaliyugarudra has ascended [from his body] and 
reached the world of Rudra'. 

648 p ANTULU i93o ; vv. 82-84, 88. 


Genesis and Development of Tantrism 

paign of world conquest. Beginning with the inscriptions of his son and succes- 
sor Karna (r. c. 1041-1071) the Kalacuri rulers of this kingdom are described as 
meditating on the feet of this Vamadeva, to whose name are prefixed the impe- 
rial epithets paramabhattarakamaharajadhirajaparamesvaraparamamahesva- 
rasrl-. A variant of these titles, samadhigatapaficamahasabdaparamabhatta- 
rakamaharajadhirajaparamesvara-, is found in Nolambavadi records attached 
to the names of two other Saiddhantika Gurus, namely Brahmasiva in an in- 
scription of c. 870 and Varunasiva in one of 936. Similarly, but more modestly, 
an inscription of 1331 on a step-well in the vicinity of the Acalesvara temple 
on Mt. Abu tells us that it was constructed during the victorious reign of the 
great ascetic rajasri- Sarvesvara during the victorious reign of the ruler rajasri- 
Tejahsimha of Candravati. 649 


For these imperial and royal titles attached to the names of Saiva Gurus see D.C. 
Sircar in EI 30:10, pp. 46-51. There he refutes the claims expressed by V.V. Mi- 
rashi in EI 27:29. These are (1) that Vamadeva is a king Vamarajadeva [seen by 
him alonel in the Saugor inscription of Sankaragana, which has been assigned on 
palaeographic grounds to the eight century, (2) that this king should be assigned 
to the second half of the seventh century, and (3) that the references in inscrip- 
tions of the later Kalacuris to these king's devotion to [the memory] of Vamadeva, 
should be referred to this much earlier monarch as the founder of their dynasty. 
Sircar removes Vamadeva from the Saugor inscription, reading -vdvardja- rather 
than -vdmardja- and citing other examples of vdva- or bdva- in inscriptions, and 
then cites these examples of imperial or royal epithets bestowed on Saiva Gurus to 
counter Mirashi's argument that their being prefixed to the name of Vamadeva 
proves that he was a king not a Guru. I side with SlRCAR. His view has the 
great strength that it accords (1) with the testimony of the Malkapuram inscription 
of 1261/2, which, referring to Vamasambhu as the third Guru in succession after 
Sadbhavasambhu, the first pontiff of the Matha at Golagi, reports that the Kalacuri 
kings were being praised (prasamsyante) [in their Prasastis] up to the present as 
worshippers of his feet (Pantulu 1930, v. 28: atha nrpasekharamdldldlitapddo 
'tra vdmasambhur abhut | adydpi kalacurisd yaccarandrddhakdh prasamsyante) — 
in the inscriptions of the Kalacuris of Tripuri from Karna onwards they are said 
to be -vdmadevapdddnudhydta — , (2) with the fact that there is no reference to 
a king Vamadeva in any of the inscriptions of those kings, and (3) with the fact 
that the source which reports the Kalacuri king's bestowing his rdjalaksmi on 
Vamadeva refers to the latter as an ascetic (sdhasikas tapasvine vdmadevandmne 
nijardjalaksmlm gurudaksindyai dattvd sarvdm bhumim jetum prasthitavdn). SlR- 
CAR convincingly identifies the Kalacuri king here called Sahasika as Garigeyadeva 
on the grounds that the latter was both an illustrious conqueror and known as 
Sahasarika 'he who has the cognomen Sahasa'. This source, cited by Sircar (EI 
30:10, p. 50), is a paraphrase in Jonaraja's commentary on the Prthivirdjavijaya of 
Jayanaka of a verse of that work now lost in a lacuna. 

The inscription referring to Brahmasiva is EC 10, Srlnivasapura taluk, 
no. 27 (p. 346). SlRCAR {EI 30:10, p. 49) wrongly gives the name 

as Bhramarasiva and the page reference as 376. The relevant part 
of the inscription is: svasti samadhigatapancamahdsabda pallavdnvaya 
sriprthivivallava pallavdnvayakulatilakam srimat-nolambddhirdjar prthivirdjya 


The Saiva Age 

Clearly the Saiva Rajaguru had become a far grander figure than the king's 
brahmanical chaplain, the Rajapurohita, who was tied to the service of a single 
king and was unambiguously his subordinate. Yet, it appears that the Saivas did 
not rest with this but also sought to encroach on the territory of that lesser office. 
For the Netratantra shows the existence of a further class of Saiva officiants who 
were to function in almost all the areas traditionally reserved for that officiant: 
the performance of the king's recurrent duties to worship the various deities 
on the days assigned to them, to celebrate the major annual royal festivals of 
the Indrotsava and Mahanavami, to protect the royal family through rites to 
ward off ills, to restore them to health after illness, to ward off or counter the 
assaults of dangerous supernaturals, to empower through lustration (nirajanam) 
the king's elephants, horses and weapons of war, and to protect the king with 
apotropaic rites before he eats, sleeps, and engages in his regular practice of 
martial skills. 650 

We see here one of several instances in which the Saivas used their author- 
ity to colonize downwards, producing modifications of their ritual procedures for 
this purpose. These adapations inevitably entailed loss of status for those that 
implemented them, but we should understand that this did not affect those at 
the summit of the clerical hierarchy, the king-like Rajagurus, but only the hum- 
bler clones that extended their authority into domains that those Gurus would 
not deign to enter. 

geye svasti samadhigatapahcamahasabda paramabhatta[ra]ka mahardja.dhira.ja 

paramesvara ataniya mata . . pana 

nvita sivasastratapovananuraga sripadhivalagramavirnirggata bha- 
gavatpadaikasarana srlmat-brahmasivacaryyan. The inscription referring to 
Varunasiva (Vanmasivabhatara) is SII 9, 1:24 (ARE 759 of 1916) from Gu- 
nimorabagalu in the Anantapur District. It speaks of him as the pon- 
tiff of the Nonambesvarara temple, as the ruler of Palivalubalu, and as the 
Mahasamantadhipati, that is to say, as a feudatory of the highest rank, of king 
Biranolamba Annayyadeva of the Nolamba-Pallava dynasty (= Anniga, r. c. 932- 
940). The Nonambesvarara is probably the imposing temple at Hemavati now 
known as Doddesvara (Cohen 1989, p. 50, and p. 63, note 36). He is also men- 
tioned in an inscription on the Mandapa of the Doddesvara temple, which gives 
the information that he was the disciple of Rudrasivacarya. On Varunasiva see 
Cohen 1998, pp. 24, 35, and 41-42, who plausibly concludes that he was Anniga's 
Rajaguru. The initiation-names Brahmasiva, Varunasiva, and Rudrasiva reveal 
that these Gurus were Saiddhantikas. The relevant portion of the inscription from 
Mt. Abu has been published by Sircar within this discussion (EI 30:10, p. 48). 
650 rp^g p ur p 0ge; d a te, and provenance of the Netratantra are the subject of SANDER- 
SON 2005b. 


Genesis and Development of Tantrism 

Saivism and the Royal Temple 

The second element of the early medieval process to which I have drawn 
attention is the proliferation of land-owning temples. All but the most ephemeral 
sovereigns during this period, both in the subcontinent and in Southeast Asia, 
gave material form to the legitimacy and solidity of their power by building grand 
temples in which images of their chosen God were installed, animated, named 
after themselves (svandmna), and endowed with land and officiants to support 
their cult. As we have seen, the great majority of these temples enshrined Siva 
[in the form of a Lihga]. 

The Saivas of the Mantramarga provided specialized officiants and rituals 
to establish these Sivas, developing in course of time a secondary body of scrip- 
tural authorities, the Pratisthatantras, devoted exclusively to this domain, set- 
ting out the rituals of installation (pratistha) and defining the norms for the form 
of the Liriga, the iconography of ancillary images, and the architectural design 
of the various temple types. 651 Moroever, they asserted the principle that the 
Saiva Sthapaka, the specialist who performs these rituals, is competent not only 
in the Saiva domain but also on all the levels that the Saivas ranked below it. 
Thus they claimed that he is empowered to officiate in the construction and con- 
secration of non-Saiva deities such as Visnu following the Pancaratra. 652 This 

651 None of the early works of this class have been published. Those known 
to learned authors before the end of the eleventh century and surviving in 
manuscripts are the Mayasamgraha, not to be confused with the published Maya- 
mata, a later south-Indian work, the Pihgalamata, the Mohacudottara , and the 
Devydmata, which declares itself the pratisthatantram of the Nisvasa. Four other 
works of this type, not known to have survived, are cited by the Kashmirian 
Vidyakantha around the beginning of the eleventh century in his commentary 
on the Mayasamgraha: the Pratisthaparamesvara, the Nandikesvaramata, the 
Paitamaha, and the Pratisthasamuccaya, the last of which was probably a Paddhati 
rather than a scripture. On all these texts see SANDERSON 2005a, pp. 440-442. 

652 See, e.g., Brhatkalottara , B f. 108v4: bauddhavaisnavapancarthe saurakalamukha- 
disu | saivah sarvadhikari syan na saive 'mi kathamcana 'The Saiva [Guru] 
has competence that extends into all [religious systems], the Buddhist, Vaisnava, 
Pancartha[-Pasupata], Saura, Kalamukha, and others; but [Gurus of] those have 
absolutely no competence to act in the Saiva [system]'; Kdmika, Purvabhaga 
1.121c-126, on the authority of the Sivabrahmanas, the married Saiva brahmins 
who alone were competent to officiate for others: saivah sarvadhikari syat sva- 
kiye ca paratra ca || 122 saivah sarvesu kurvanti ye grhastha dvijottamah \ yamale 
mdtrtantre ca kapale pancaratrake || 123 bauddhe carhamate caiva lakule vai- 
dike 'pi ca | anyesv api ca margesu tattacchastraih svasastratah || 124 saivah ku- 
rvanti diksadyam tallihgasthapanadikam | mukhyatvad iha saivasya mukhamaha- 
tmyato 'pi ca \\ 125 adhikaro 'sty sarvatra nanyesam sivadarsane \ tasmat para- 
rtham atmartham sthapanam yajanam tatha || 126 sivaviprena kartavyam anyesam 
svartham eva hi | parartham api kuryac eel Hobhena (em. : lopena Ed.) nrpates 
tatha | tadrastrasya ca nasah syad acirena na samsayah 'The Saiva is competent in 


The Saiva Age 

universalization of their authority, which is backed by learned theory of the re- 
lation of the Saiva with the other bodies of scriptural injunction, seems not have 
been merely theoretical. For the Saiva Paddhati literature contains instructions 
for the consecration of Visnus, as we have seen in the case of the Paddhati of 
Somasambhu, 653 and Vaisnava sources protest at this encroachment, insisting 
that images of Visnu installed by Saivas should be reconsecrated. 654 

all [systems], both his own and others. Married Saivas, the foremost of brahmins, 
can officiate in all [the systems, namely] the Yamala and Matrtantra, the Kapalika, 
the Pancaratra, the Buddhist, the Jaina, the Lakula, the Vaidika, and yet others, 
using the scriptures of these systems in accordance with their own. [Such] Saiva[ 
brahmin]s perform initiations and the like, the installation of images, and so forth 
[in these other systems], because the teaching of Siva is superior [to all others] and 
because the mouth [of Purusa] has been glorified [in the Purusasukta as the part of 
his body from which the brahmins, as the highest caste-class, were created]. [The 
Sivabrahmana] is competent to act in all [systems], but not others in the teaching 
of Siva. Therefore the Sivabrahmana [alone] may worship and install [images] both 
for others and himself. Others may act only for themselves. If out of greed [anyone 
other than a Sivabrahmana] performs rituals for the benefit of others[, thus usurp- 
ing the exclusive right of the Sivabrahmanas], then without doubt both the king 
and his kingdom will swiftly be destroyed'. 

653 See Somasambhupaddhati vol. 4, pp. 294-311 (visnusthdpanavidhih). 

654 In his Pancardtraraksd (pp. 26-27) Vedantadesika, the influential Srivaisnava 
of the fourteenth century (EI 13, p. 222), quotes a passage from the Saiva 
Kdranatantra that is more or less identical with 1.121c-124 of the passage of 
the Kdmika, Purvabhdga cited above, and after asserting that it is inadmissi- 
ble as evidence because all Saiva Tantras are condemned by Vedic authorities 
quotes a passage from the south-Indian Pancaratrika Pddmasamhitd (Carydpdda 
19.128b-130) to the effect that if a Visnu has been installed with the system of 
the Saivas it must be re-installed following the system of the Pancaratra and pu- 
rified by bathing with a thousand vases. See also Visvaksenasamhitd 39.283- 
285: sthdpite raudramdrgena pujyamdne dine dine \ hitvd raudravidhdnam tu 
sarvesdm hitakdmyayd || grdmavrddhikaram punyam rdjabhusuravardhanam | 
tasmdt sarvaprayatnena hitvd raudram tu tatksandt || sthdpayet sdttvatendtha vid- 
hind pujayed dharim | tasmdt sarvaprayatnena na kurydt tantrasamkaram 'If [a 
Visnu] has been installed following the Saiva procedure and is in daily worship [fol- 
lowing the same] then, desiring the welfare of all, one should abandon the Saiva 
procedure and [adopt] the holy [Vaisnava procedure] that will cause the village, the 
king, and the brahmins to prosper. Therefore one should abandon the Saiva rites 
immediately and scrupulously re-install the Visnu with the Pancaratrika ritual and 
and worship it [with the same thereafter]. So one must take great care to avoid 
[this] contamination of the [Saiva and Vaisnava] systems of worship'; 39.305-306: 
jdtisamkaranenaiva jagac canddlatdm vrajet | tantrasamkaranenaiva rdjardstram 
vinasyati \\ rdstram sariram rdjnas tu rdjdjivah sa ucyate | rdstraksaye ksayo rdjnah 
tasmdd raksyam dvayam budhaih 'People become untouchables through the con- 
tamination of castes. Through the contamination of the systems of worship the king 
and kingdom are destroyed. [The scriptures] teach that the kingdom is the body 
and the king its soul. [So] when the kingdom is destroyed, so is the king. The wise, 
therefore, should guard both [by preventing the encroachment of the Saivas into the 
Pancaratrika domain]'. Saiva ritual is called raudra- in the first of these passages 


Genesis and Development of Tantrism 

The involvement of the Saivas of the Mantramarga in the temple cult 
covered in early Saiva scriptural sources and all the early Paddhatis up to at 
least the twelfth century does not extend beyond the performing of the rituals 
necessary to initiate the cult by consecrating the images and the temples that 
house them. The texts are silent on the nature of the worship that would be 
performed before those images once the Saiva Guru had completed his task. It 
would appear, therefore, that the temple worship was in the hands of officiants 
of a different kind. However, the texts lagged behind reality in this regard. For 
at some point, well before the Saiva literature was prepared to register this fact, 
there were Saivas of the Mantramarga working as the priests that performed 
the regular rituals in the Saiva temples. The new practice is first attested in 
the Far South in the late seventh century. We learn from a grant of the Pallava 
Paramesvaravarman I (r. c. 655-960) that a certain Anantasivacarya, whose 
name makes it very probable that he was an initiated Saiddhantika officiant, 655 
was appointed as the priest with hereditary rights to perform the ritual of 
worship (devakarma) in the temple of Siva Vidyavinitapallavaparamesvara 
established with his name by the Pallava king Paramesvaravarman I alias 
Vidyavinita. 656 

The persistent disjunction during this period between what was prescribed 
for Saivas and what was being done by some of them is due, I propose, to the 
fact that functioning as a priest in a temple, and therefore living off the endow- 
ment of the deity in return for one's work, carried a loss of status with which 
the older tradition was unwilling to be associated. According to brahmanical 
sources any brahmin who persists in such work for three years is considered to 

in keeping with the mildly disparaging south- Indian Vaisnava practice of referring 
to Siva as Rudra. Cf. the expression rudrakalyupajivakah cited here, p. 278 and the 
rule of the Sandilyasmrti quoted by Vedantadesika in his Pancaratraraksa (p. 62) 
that Vaisnavas should keep far away from temples of Buddha, Rudra, and the like 
(buddharudrddivasatim smasanam savam eva ca | atavim rajadhanlm ca duratah 

655 Saiddhantika Saiva initiated brahmins have initiation-names (diksdndma) that end 
in -siva (with -sambhu or, less commonly, -isvara/-isa or -sarikara as synonyms) 
as the second of their two components, and those of these who have been conse- 
crated to officiate by receiving the dcdrydbhisekah are referred to as N-sivacarya, a 
practice that has continued into modern times. Other Anantasivacaryas are the 
author of the Siddhdntasdrdvallvydkhyd, one of the Sivacaryas, probably 95 in 
all, among 108 12th-century labelled images at Darasuram in Tamilnadu (SRINI- 
vasan 1987, vol. 1:17, no. 60), and one mentioned in an inscription of 1571 at the 
Vatar any es vara temple at Tiruvalarigadu (ARE 497 of 1906 [Appendix B: stone in- 
scriptions copied in 19051) as a disciple of Ponnambala Dharmasivacarya and Guru 
of Immadi Dharmasivacarya. 

656 The Kuram plates of Paramesvaravarman I (r.c. 655-90): MAHALINGAM 1998:46, 
11. 55-57 (Sanskrit) and 11. 84-88 (Tamil). 


The Saiva Age 

have lost his brahmin status and is then known as a Devalaka. 657 He is de- 
scribed as an upabrahmanah 'a sub-brahmin' or, even more disparagingly, as a 
brahmanacandalah 'a brahmin untouchable'; 658 and this loss of status is con- 
firmed in modern times in the way that the Smartha brahmins, the dominant 
community in Tamiladu have viewed the Adisaiva community that provides the 
priests who after undergoing Saiddhantika Saiva initiation (diksa) and consecra- 
tion as Acaryas (acaryabhisekah) perform the worship in the Siva temples of the 
region. They were forbidden to live in brahmin streets and the Smarthas would 
not intermarry or interdine with them. 659 The Adisaivas, as one might expect, re- 
sisted this condemnation, arguing in their scriptural productions and in learned 
exegesis that it applies only to brahmins other than members of their endoga- 
mous community, more precisely that the three-year rule applies to Saiva initi- 
ates other than themselves. Strengthening the brahmanical position they held 
that Siva has ruled that ordinary, uninitiated brahmins who work as temple- 
priests will forfeit their status after only six months. 660 As modern practice 

657 Yamuna, Agamapramanya, pp. 15-16: tatha ca devalah "devakosopajivi yah sa de- 
valaka ucyate" iti | tatha "vrttyartham pujayed devam trini varsani yo dvijah \ sa vai 
devalako nama sarvakarmasu garhitah" iti 'And Devala [teaches]: "One who lives 
off the wealth of a god is called a Devalaka", and: "Any brahmin who does the wor- 
ship of a god for his living for three years is called a Devalaka, and is condemned in 
all rites'". By 'condemned in all rites' the text means that such a brahmin must not 
be chosen as an officiant in any brahmanical ritual or invited as a participant in a 

658 ^ r j c ited in Agamapramanya, p. 16: tatha ca visadataram amlsam 
evopabrahmanyam varnayaty atrih: "ahvayaka devalakah kalpadevalaka 
ganabhogadevalaka bhagavatavrttir iti caturthah. eta upabrahmanah" iti 'And 
Atri makes it absolutely clear that it is those that are sub-brahmins, when he 
says: "Couriers, Devalakas, Kalpadevalakas, Ganabhogadevalakas, and fourth, 
he who lives by being a Bhagavata: these are sub-brahmins'"; and Mahabharata 
12.77.8: ahvayaka devalaka naksatragrdmaydjakdh ete brahmanacandala 
mahapathikapahcamah 'All the following are brahmin untouchables: couriers, 
temple-priests, those who perform worship to the asterisms, those who perform 
worship on behalf of a whole village, and, fifth, those who undertake long journeys'. 

659 See Thurston 1909, p. 51, and Fuller 1984, pp. 49-71. The Dikshitars, the 
priests of Siva at Cidambaram, rank above the Adisaivas, probably because they 
are the trustees of their temple; but they are still considered inferior to non-priestly 
brahmins; see Fuller 1984, p. 192, n. 3. 

660 Vedajnana II, Atmarthapujapaddhati A, p. 123 and B, p. 99, quoting the Viratantra 
and the Samtanatantra: viratantre "bhrtyartham sarvadakdlam adisaivah sivam 
yajet | tac ca svadharmanusthanam na dosaya prakalpate || adlksitas caturvedl siva- 
lihgam na samsprset | diksitas cdpi yo vipro bhrtyartham tu na pujayet || atmartha- 
pujam kuryat *pardrtham naiva (A:pararthan caiva B) pujayet" | samtane "adlksito 
'pi yo viprah sanmasam tu sivam sprset | so 'pi devalakah proktah sa narho deva- 
pujane | diksitas cdpi yo vipro *bhrtyartham (em. \pratyartham A : bhrtydnced B) 
vatsaratrayat \ pujayed yadi devesam so 'pi devalako bhaved" iti 'Viratantra: An 
Adisaiva may worship Siva for a living permanently; and that, since it is his reli- 


Genesis and Development of Tantrism 

reveals, this counterargument had no effect on the Smartha majority; and, in- 
deed, it is obvious that its real purpose was rather to defend their professional 
rights against encroachment by others, rights that they took care to write into 
their scriptures. 661 For, no doubt in consequence of the efflorescence of the Saiva 
temple cult under the Cola emperors, we find a new wave of Saiva scriptures 
appearing in the South, in which the ceremonial life of the temple and the duties 
and rights of its priests are regulated, and, indeed, form their principal subject 
matter. Citations from the majority of the scriptural texts of this kind do not 
appear before the works of Vedajnanaguru II, composed during the second half 

gious duty, cannot be sinful [for him]. An uninitiated [brahmin], [even if he is one] 
who knows [all] four Vedas, may not [even] touch the Lihga of Siva; and even a 
brahmin who has been initiated may not worship [it] for a living [unless he is an 
Adisaiva]. He should worship [Siva] for his own benefit [as a private individual]. 
He may not also worship him for the benefit of others [as a priest in the temple]. 
Samtana: If an uninitiated brahmin has physical contact with a Siva [installed 
in a fixed Lihga in a temple] for six months he is called a Devalaka and is dis- 
qualified from offering worship to [any] deity [thereafter]. Even an initiated brah- 
min becomes a Devalaka if he [is not an Adisaiva but] worships Siva for a living, 
once three years [of his doing so] have passed'; and Kacchapesvarasivacarya, Kriya- 
kramadyotikavyakhya, p. 80, 11. 4—7, quoting the Viratantra: adiksitas caturvedi na 
sprsen napi cdrcayet | bhrtyartham paramesanam dlksavirahita janah | *sanmasad 
ydnti (em. : sanmasavyanti Cod.) patityam te ca devalakah smrtah || trini varsani 
bhrtyartham sthiralihge *hi diksitah (em. :hy adiksitah Cod.) | pujayed yadi *vipras 
(corr. wipras Cod.) tu sa vai devalako bhaved iti An uninitiated [brahmin], [even if 
he is one] who knows [all] four Vedas, may not touch and worship Siva for a living. 
The uninitiated fall from their caste after six months [if they do so]. It is they that 
are known as Devalakas. If an initiated brahmin [who is not an Adisaiva] performs 
the worship [of Siva] in a fixed Lihga for a living for three yearsf, that is to say, as 
a priest serving in a temple,] then he [too] will become a Devalaka'. In the older, 
north-Indian literature the Prayascittapatala of the Dvadasasahasra Svacchanda, 
quoted by Hrdayasiva in his Prdyascittasamuccaya, f. 92v3-4, defines Devalakas 
when considering the matter of contamination by them, as those who as priests 
(bhojakah) live off the Moon-god, Brahma, the Sun-god, Skanda, Visnu, the God- 
dess, or the Mothers: somabrahmaraviskandavisnudevyas ca mdtarah | upajivanti 
ye devi pujayitvd tu bhojakah | te vai devalakas tesam prayascittam vadamy aham. 
The omission of Siva from this list implies that it is only the priests of other gods 
that fall from caste. Likewise, defending the Paiicaratrika priests of Visnu's tem- 
ples against the same consequence, Yamuna argued, citing Vyasa, that it is only 
those who earn their live off Rudra (i.e. Siva) and Kali by serving as their priests 
that become Devalakas (bhaved devalako yo vai rudrakalyupajivakah): Vaisnava 
temple-priests do not become Devalakas, because they have been consecrated for 
their work by initiation. See Yamuna, Agamaprdmanya, pp. 15-17 (the accusation), 
and pp. 156-157 (the rebuttal). 
1 See, for example, the Kamika cited here p. 274, the Vira and Raurava cited in 
Brunner 1964, p. 468, n. 11, and the Yogaja, Cintya, Vira, Samtana, and other 
Agamas cited by Vedajnanaguru II in his Atmarthapujapaddhati A, pp. 121-123 
and B, pp. 97-99. 


The Saiva Age 

of the sixteenth century. 662 But some others are already being cited in the thir- 
teenth, and one in the twelfth. 663 

Here too, of course, the royal connection is maintained and carefully nur- 
tured. Thus the ceremonial repertoire of these temples included special rituals 
for the king's protection (rajaraksa); 664 and temple festivals (utsavah) were often 
timed to coincide with the day of his natal asterism or of that of a member of 
his family 665 Indeed the texts place a great emphasis on the connection between 
the temple and the welfare of the ruler and his kingdom, warning repeatedly 
that while the proper maintenance of the temple and its ceremonies will bene- 
fit both, deviations or neglect will have dire consequences for them. This duty 
to maintain the status quo naturally included that of recognizing the exclusive 
hereditary rights of the members of this priestly community 666 

The Adisaivas are the only endogamous community of Saiddhantika Saiva 
temple-priests for which we have evidence and they seem not to have operated 
beyond south India. But it seems likely that there were parallel developments in 
other parts of the subcontinent, evidence of which has been lost or not yet come 

662 These scriptures that first appear in the works of Vedajnanaguru are the Amsumat, 
the Ajita, the Kdsmiratantra , the Cintyavisva I Cintyavisvasdddkhya, theDTpta, the 
Devikdlottara, the Bhima, the Makuta, the Mukhabimba, the Yogaja, the Raurava, 
the Vijaya (IVijayottara), the Vidvesana, the Vira, the Samtdna, the Sahasra, the 
Siddha, the Suksma, and the Skandakdlottara. The works of Vedajnanaguru in 
which they are cited are the Atmdrthapujdpaddhati, Dlksddarsa, and Saivdgama- 
paribhdsdmanjari. For his date see Dagens 1979, pp. 6-7. 

663 rp^g ex ^ an ^ Kdmika is perhaps the first work of this kind to be cited in a date- 
able work. Substantial passages found in it are quoted without attribution in 
the Jhdnaratndvall of Jnanasiva, a teacher of Trilocanasiva and therefore a near 
contemporary of Aghorasiva, who completed his Kriydkramadyotikd in 1157. The 
next earliest known work in which there are citations from such scriptures is the 
Sivapujdstavavydkhyd composed by a nameless author in the thirteenth century, 
probably in its second half. This date follows from the fact that he identifies himself 
as the great-great-grandson of the same Trilocanasiva. He cites the Kdrana, the 
Acintya, the Suprabheda, the south-Indian Pauskara, and the Vdtulasuddhdkhya. 
I derive this information concerning the citations in the Sivapujdstavavydkhyd and 
Jnanasiva's unattributed citations of the Kdmika from a lecture given by Dr. Do- 
minic Goodall in the Early Tantra Workshop held in Kathmandu in September 
2008. For the relationships between Aghorasiva, Jnanasiva, and Trilocanasiva see 
Goodall 2000 and for confirmation of the date of Aghorasiva's Kriydkramadyotikd 
see Goodall 1998, pp. xiii-xvii, fn. 24. No Sanskrit Saiddhantika works have yet 
been identified which can be dated within the period of three centuries between the 
author of the Sivapujdstavavydkhyd and Vedajnanaguru II. 

664 Chapters devoted to this protective temple ritual for the king are found in such 
south-Indian Saiva texts as the Suksmdgama (pp. 290-297: rdjaraksdvidhih), and 
the Diptdgama (pp. 211— 215: rdjaraksdvidhipatalah). 

665 See Davis and Orr 2007, p. 91, for epigraphical evidence of such arrangements. 

666 See, for example, the passage of the Kdmika cited above, p. 274. 


Genesis and Development of Tantrism 
to light. 667 

Saivism and New Settlements 

The early Saiva Pratisthatantras show that the authority of the Saiva 
Sthapaka was to extend to the creation of the palaces of their kings. Among the 
early Pratisthatantras the Mayasamgraha, Mohacudottara, and Pihgalamata, 

667 Against the view that the Adisaiva caste is peculiar to Tamil Nadu one might 
cite the fact that the Adisaivas are mentioned the Somasambhupaddhati, a work 
composed in the eleventh century far to the north (at the end of the Pavitrd- 
rohanavidhi): pancayojanasamsthe 'pi pavitram gurusamnidhau \ kurvita vidhi- 
ndnena labhate vanchitam phalam | sarvam vai tv adisaivanam diksitdndm 
sivoditam | paropakarasilena srimata somasambhuna | kriydkdndakramdvalydm 
pavitrakavidhih krtah. However, the line is not in the edition based on Kash- 
mirian manuscripts (see Karmakdndakramdvali vv. 494c-496b: pancayojana- 
samsthe 'pi pavitram gurusamnidhau \\ kurvita vidhindnena labhate vanchitam 
phalam \ adhitasivasastrena krto 'yam somasambhuna || karmakdndakramdvalydm 
pavitrakavidhih sphutah) nor in the Nepalese transmission (see Kriydkdndakramd- 
vali f. 22v4— 5: pancayojanasamstho 'pi pavitram gurusamnidhau \ kurvita vidhina- 
nena labhate vanchitam phalam \ paropakarasilena srimata somasambhuna \ kriya- 
kandakramavalyam pavitrakavidhih krtah). It is found only in Brunner's edition 
and the Devakottai edition, which her edition reproduces here. It rests, therefore, 
exclusively on the evidence of Grantha manuscripts from the south. Evidently, then, 
one must suspect that the line has been interpolated in Tamil Nadu by a redactor in 
the Adisaiva community Its lack of intelligible connection with what precedes and 
follows strengthens this suspicion. 

I have not seen the term Adisaiva in any inscription. There the officiants of 
the Siva temples are always termed sivabrdhmanah or sivadvijah. That term 
first occurs to my knowledge c. A.D. 863 in an inscription of Pallava Nandivar- 
man III, from Tiruvallam in North Arcot (Mahalingam 1988:132). Concerning 
a grant to the temple of Paramesvara at Tikkalivallam it specifies that 500 kadi 
of paddy are for the Sivabrahmanas who offer worship and services in the sanc- 
tum (dr[d]di[t]t-updsarikkum [sivajbrdhmanarkku) (11. 25-26). Thereafter the term 
is commonplace. But it is clear that it is the group known as the Adisaivas that 
is intended, because in these inscriptions when Sivabrahmanas are named their 
Gotras are sometimes given and these are those of the Adisaivas as attested both 
by their prescriptive texts and among their modern descendants, namely Kausika, 
Kasyapa, Bharadvaja, Gautama, Atreya, Agastya, and Parasara. See, e.g., SII 
3:41 (Kasyapa), 55 (Kausika), 58 (Kausika), 209 (Kausika, Kasyapa, Kausika); 
SII 12:197 (Agastya); SII 17:152 (Bharadvaja), 157 (Bharadvaja), 160 and 161 
(Gautama), 162 and 163 (Bharadvaja), 165 (Gautama, Parasara), 203 (Atreya, 
Bharadvaja), and 730 (Kasyapa); EC 3, Sr:44 (Gautama); EC 10, Kl:106a (Kausika), 
106d (Gautama), 107 (Kausika), 187 (Kausika, Kasyapa); EC 10, Bp:29, 32, 35a, 
and 37a (all Gautama, Bharadvaja). Six of the seven, minus Agastya, are listed in 
the Samtdna as cited in the Atmdrthapujdpaddhati A, p. 125. Five of them, lacking 
Atreya and Parasara, are listed in Svdyambhuva, p. 14 (Acdryalaksanapatala 94c- 
95b). This evidence accords with contemporary testimony. According to the data 
collected by Fuller (1984, p. 28) the Adisaiva priests of the Mmaksisundaresvara 
temple in Madurai belong to the Kasyapa, Kausika, Bharadvaja, Gautama, and 
Atreya Gotras. 


The Saiva Age 

all prescribe the layout of the royal palace in detail, the latter two distinguishing 
between different classes, the highest being that of a paramount sovereign 
or Maharajadhiraja; 668 and in the first two works the design prescribed in- 
cludes a section of the palace reserved for teachers of the Saiva Mantramarga 
(mantrinah, mantramdrgopadesinah). 669 But the layout of the palace taught 
in these Pratisthatantras is only part of the layout for an urban settlement 
to be established by the king around the palace, complete with markets and 
segregated areas for the dwellings of the various castes and artisans, with 
instructions for the size and plan of these dwellings determined by caste sta- 
tus. 670 The founding of such royal towns is not explicitly enjoined in the Saivas' 
ritual manuals. That is to say that no ritual of nagarapratisthd was envisaged. 
The Sthapaka was engaged, it seems, only for the choice and consecration of the 
site (vdstupuja) and his instructions followed for the layout of the buildings to be 
constructed upon it. Nonetheless, we see the Saivas involving themselves in one 
aspect of the third of the elements of medieval process that I have listed, namely 
the creation of new urban settlements from above. The epigraphical record and 
Kalhana's history of Kashmir demonstrate that any king of substance felt it 
encumbent on him to demonstrate his sovereignty not only by the building of 
temples but also by the creation of new urban settlements (puram), which, like 
the deities he established, were generally named after him. 671 

One of the early Pratisthatantras, the Devydmata, devotes its 66th chapter 

668 rp^g i a y OU £ f th e r0 y a i palace is prescribed in Mayasamgraha ff. 33v-34r 
(5.188-199), Mohacudottara ff. 20v-22r (4.245c-281), and Pihgalamata ff. 74r-75v 

669 Mayasamgraha ff. 33v-34r (5.191-193b): vitathe mantrinam dhama 
sarvastrani grhaksate | antahpuram yamapade gandharve gatrsamsrayam 
|| bhrhge senapatisthanam mrganabhyadikam mrge | paitre saucagrham 
cdtra tambuladivyapasrayam \\ avarodhavadhusthanam sugrive tu tato 
nyaset; Mohacudottara 4.257c-258b: vitathe mantrinam sthanam 
mantramargopadesinam || sastram antahpuram gatr kasturi saucavesma 
ca | tambulasamgrahah strinam *palakan (em. :pacakan Cod.) striniyamakan. 

670 Mayasamgraha ff. 34v-35r (5.209-216); Mohacudottara f.21vl-6 (4.270-275b); 
Pihgalamata ff. 75v-76r (10.181-194). 

671 This practice was followed both throughout the subcontinent and in Southeast 
Asia, as the following examples demonstrate: in Kashmir Pravarasena IFs Pravara- 
pura (Srmagar), Durlabhaka-Pratapaditya II's Pratapapura, Jayapida's Jayapura, 
Lalitaditya's Lalitapura, Avantivarman's Avantipura, Sankaravarman's Sankara- 
pura, and Didda's Diddapura, in eastern India Ramavati (Ramauti) (of Ramapala), 
Vijayapura of Vijayasena, and Laksmanavati (Lakhnauti) (of Laksmanasena), 
in the south Gangaikondacolapura, Parakesaripura, Parantakapura, Rajakesari- 
pura, Rajarajapura, Rajadityapura, Rajasrayapura, Rajendracolapura, Vikrama- 
colapura, Vikramapandyapura, Vikramasimhapura, Virarajendracolapura, Vira- 
rajendrapura, and, among the Khmers Isanapura, Bhavapura, Yasodharapura, 
Rajendrapura, and Jayendranagari. 


Genesis and Development of Tantrism 

to the layout to be followed not only in new towns but also in new villages, with 
an emphasis on the positioning of the various deities within the plan and the 
directions in which they should face. The regulations imposed show us Saiva of- 
ficiants on a purely civic level. There is nothing specifically Saiva in the layout. 
The Devyamata's chapter on iconography shows further evidence of the involve- 
ment of the Saivas in both urban and rural planning. Differentiating various 
forms of Siva in accordance with mood and number of arms it tells the Sthapaka 
which are appropriate where. 672 The same concern can be seen in the Pratistha 
sections of the South-Indian Yamalatantra texts with regard to the positioning 
and iconography of the images of Bhadrakali whose installation and cult are 
their concern. 673 

Saivism and Irrigation 

The creation of new settlements and the concomitant extension of agricul- 
ture required the provision of the means of irrigation. Rituals for the conse- 
cration (pratistha) of wells (kupah), step-wells (vapi), and reservoirs (puskarinl, 
tadagah) were already provided by the brahmanical tradition. A Vaidika proce- 
dure of the Grhya type is outlined or touched upon in a number of sources; 674 
a more elaborate, Pauranika form of the ritual, taking five days and requiring 
twenty-four priests in addition to the Sthapaka, is set out at length in the Mat- 
syapurana (58.4-56); 675 and the currency of this form is evident from the fact 
that it became the basis of further elaboration. 676 There is no trace of irriga- 

672 Devydmata f. 68r4: dvibhujo rdjadhdnydm tu pattane tu caturbhujah | tathd cdsta- 
bhujo bhadre prasastah pattane sthitah. 

673 Thus in Brahmaydmala IFP 40.1— 4b: atah param pravaksy ami pratimdlaksanam 
param | navatdlapramdnena pratimdm kdrayed budhah || 2 sildmayam lohamayam 
mrnmayam vapi kdrayet | grdme cdstabhujam vidydn nagare ca caturbhujam \\ 
3 vandntare dvibhujam vidydt parvatdgre tu sodasa | samudre dvddasam kurydt 
*jandandya (?) . . . sadbhujam || 4 tatdke dasabhujam kurydt catuspathe caturbhu- 
jam; and Brahmaydmala Triv. 3.3-8: grdme ca nagare caiva pattane rdjadhdnike 
| raksdrtham vdstavasthdnam pure vai khetakddisu || 4 sarvasddhdranam vidydd 
yathdvibhavavistaram \ bahih prakdratah kurydn mdtrsthdnam tu vdstavam \\ 
5 srestham purvottare bhdge satadanddnta'nantare | tadardhe vdtha tasydrdhe 
dasadanddntare 'pi vd || 6 some sydd vdstavam brahman mdtrndm iha codi- 
tam | purve vd pascime vapi sthdnam asya prasasyate || 7 yo me purvottare 
vapi nagaragrdmasobhitam \ daksine ketakasyoktam anyesdm prdci pascime || 8 
dgneyanairrtais caiva trttyam vdyugocaram | + + [ijttham prasamsanti ydmale 
sivabhdsite. On these south-Indian Yamala texts, the cult they teach, and their 
non-brahmin priests see SANDERSON 2007b, pp. 277-278 with footnotes 140-143. 

674 See Einoo 2002 for the details of these sources. 

67s a procedure of the Pauranika type is also taught in Asvaldyaniyagrhyaparisista 4.9 

and Hir any a 1.7.1. (ElNOO 2002, pp. 713-714). 
676 We find procedures based on the prescriptions of the Matsyapurdna in the rit- 


The Saiva Age 

tion rituals in the early Saiva scriptures, including the Pratisthatantras. But in 
due course Saiva officiants, seeking to add this important domain to their ritual 
repertoire, produced their own version. It first surfaces in our surviving evidence 
towards the end of the eleventh century, in the Paddhati of Somasambhu, 677 and 
from that source entered both later Paddhatis such as the Siddhantasekhara and 
the Atmdrthapujdpaddhati and the second wave of Saiva scriptural literature 
produced in southern India. 678 In spite of the Saivized character of these new 
rituals the underlying model is still recognizably that of the brahmanical tradi- 
tion. The Saiva elements are little more than a veneer on what it essentially a 
brahmanical procedure, marked by such distinctive features as the erecting of 
a Naga pole (nagayastih) at the centre of the excavation, the casting of metal 
images of aquatic creatures into the water, the crossing of the excavation by a 
cow followed by the patron of the rite, the making of offerings to Varuna, and 
the giving of the cow to the officiant. 679 Nor is there any attempt to attribute to 

ual literature of the priests of the Kashmirian brahmins; see *Vdpyddiprati- 
sthd, ff. 893rl4-905vl6 (Varunapratistha); ff. 906rl-907v9 (the Varunapratistha of 
Jlvana); 910rl— vl (Adityapurdne Nalakapratistha); 929v7— 931r8 (Taddkapratisthd 
and Nalakapratistha); and 931r9-931v23 (Chandogapratisthdtah Kupapratistha). 
These treatments do not appear to be distinctively Kashmirian. On the subject of 
the giving of wells and reservoirs and the Smarta/Pauranika procedures for conse- 
crating them see also Caturvargacintdmani, vol. 1 (Ddnakhanda), pp. 1001—1029. 

677 See Somasambhupaddhati, Brunner 1998, pp. 392-403 and pp. 406-411. The first 
passage sets out the ritual for the consecration of a puskarini, but adds at its end 
that it applies also for the consecration of a vapi or tatdkah. The second passage 
gives the ritual for the consecration of a kupah. A kupah is a simple well, whereas 
a vapi is a step-well, a well with a flight of steps leading down to it on one of more 
sides (kupo 'dvdrako gartavisesah baddhasopdnako 'yam vdpiti dvaitanirnayah: 
Raghunandana cited in Kane 2ii, p. 893). Such step-wells survive from the early 
medieval period, notably in Gujarat. The most splendid is no doubt the Rani ki 
Vav at Patan (Anahillapattana), the old Caulukya capital. Both a puskarini and a 
tatdkah (/taddgah) are water reservoirs. The difference appears to be one of scale 
alone, the latter being larger than the former. Kane (loc. cit. ) reports the view 
expressed by Raghunandana in his Jaldsayotsargatattva that a puskarini is from 
100 to 200 cubits in length, and a taddgah is from 200 to 800, and the view of the 
Vasisthasamhitd as quoted by Raghunandana that a puskarini is up to 400 cubits 
in length and a taddgah up to 2000. 

678 See Siddhantasekhara of Visvanatha (13th century, Benares), pp. 565-568 (11.1- 
28b); Atmdrthapujdpaddhati of Vedajfianaguru II (16th century, Cidambaram), 
A, pp. 621-629, citing from the scripture Cintyavisvasdddkhya a passage obvi- 
ously incorporated from the Somasambhupaddhati (see Brunner 1998, p. 392, 
fn. 1); 'Kriydkramadyotikd , MS transcript, pp. 344-346 (Kupapratistha); Virdgama, 
Patala 92. The section of the Somasambhupaddhati on the consecration of reser- 
voirs is also included in the Kashmirian *Vdpyddipratisthd (ff. 907vl0-908r9). 

679 See Somasambhupaddhati, Brunner 1998, pp. 397-403 (w. 8-19). Saiva ele- 
ments: the officiant recites the Pasupatastra Mantra as the patron crosses with 
the cow, makes oblations with the Aghora Mantra, instead of making an offering to 
the Vedic god Varuna may to do so to the Saiva Vamadeva, and after preparing a 


Genesis and Development of Tantrism 

the ceremony any specifically Saiva purpose or meaning. A work of public utility 
(purtam) after all is just that. 

That Saiva officiants were engaged to perform the consecration of irrigation 
works undertaken by their royal patrons seems very likely. No inscription known 
to me records any such ritual, but then no inscription to my knowledge conveys 
information about any religious ceremonies that accompanied the inauguration 
of reservoirs and other such works. It is even more probable that the Saiva ver- 
sion of the ritual would have been performed when Saiva Gurus undertook such 
constructions in their own right. We have seen above that inscriptions record the 
creation of reservoirs by Vimalasiva, Murtisiva, Prabodhasiva, Patarigasiva, and 

Saivism and Social Integration 

The fifth and last respect in which Saivism can be seen to have played an 
active role is that of the assimilation of the communities that were caught up 
in the extension of the reach of the state that characterizes this period. For the 
Saiddhantikas opened initiation to candidates from all four caste-classes, 680 in- 
cluding the Sudras or at least the Sacchudras or 'Pure Sudras', those, that is, who 
had already succumbed to the values of brahmanical society to the extent that 
they had abjured alcohol, 681 a move that both promoted the penetration of these 

porridge (caruh) with the Mantra of either makes the full oblation with the porridge 
using the Mantra of Siva. 

680 Vaktrasambhu, Mrgendrapaddhativydkhyd, p. 188: srlmatpauskare 'pi: brdhmandh 
ksatriyd vaisydh sudras caiva striyas tathd | *jaddndhabadhird (em. : jaldndndha- 
ttrako Cod.) mukd diksydh *saktipracoditdh (sakti em. : sakttha Cod.) 'And in the 
Pauskarafpdramesvara]: Brahmins, Ksatriyas, Vaisyas, Sudras, women, imbeciles, 
the blind, the deaf, and the dumb: all should be initiated if they have been in- 
spired by [Siva's] power'; and Raurava quoted by Bhatta Ramakantha on Matahga- 
pdramesvara, Kriydpdda 5.93 in support of the view that candidates for initiation 
should be brought before the Mandala in the order of their castes: yad uktam 
srlmadrauravddau: brdhmandn ksatriydn vaisydn sudrdms caiva striyas tathd 'As 
has been taught in such scriptures as the Raurava: brahmins, Ksatriyas, Vaisyas, 
Sudras, and women'. 

681 Pardkhya cited by Trilocanasiva in Prdyascittasamuccaya, p. 141: yad uktam 
srimatpardkhye: kdryd diksdpi sarvesdm *tacchaktividhiyogindm (tacchakti 
corr. : tacchaktir Cod.) | traydndm api varndndm na tu sudrdntyajdtisu | amadyapds 
tu ye sudrdh saivdcdrakriy*ddardh (corr. :ddirdh Cod.) | sivabhaktds *ca (corr. :cai 
Cod.) tesdm sd dlksd *kdrydnyaihd na hlti (em. : kdryannyathdnuhlti Cod.) 'As has 
been taught in the Pardkhya: 'Initiation should be done for all who have received 
the action of [the descent of] his power, for all three caste-classes but not for [ordi- 
nary] Sudras and the lowest-born [below them]. One may initiate Sudras, but only 
those who do not drink alcoholic liquor, who revere the disciplines and rites taught 
by Siva, and are devoted to Siva themselves'. 


The Saiva Age 

values and enabled the integration of the landowning agriculturalists, classed 
as Sacchudras, that were dominant in the countryside both within and beyond 
the core territories of these expanding states. It thus provided a means of artic- 
ulating a social unity that transcended the rigid exclusions of the brahmanical 
social order. Nor did it allow non-brahmins only to be initiated. More crucially 
it sanctioned their appointment as Acaryas, restricting this licence only by re- 
quiring that persons could officiate for persons of none but their own or inferior 
caste-classes. Thus a brahmin could teach, initiate, and perform ceremonies of 
installation only for brahmins, Ksatriyas, Vaisyas and Sudras, a Ksatriya only 
for Ksatriyas, Vaisyas and Sudras, a Vaisya only for Vaisyas and Sudras, and a 
Sudra only for others of his caste-class. 682 The key groups here appear to have 
been the first and the last. For there is little evidence of the presence of Vaisya 
traders in Saivism, and though, as we have seen, Ksatriya rulers were commonly 
Saiva initiates, their social status and function were obviously incompatible with 
pontifical office. The core social structure here is one of brahmin Gurus initiat- 
ing other brahmins, Ksatriyas rulers, and perhaps on occasion members of lower 
castes, and of Sudra Gurus initiating both other Sudras and the powerful in their 
communities, who though ksatriya-like in their local authority 683 were nonethe- 
less formally of the same caste-class as their initiators. The Sastric formulation 

: Kirana f. [60]v2-3 (38.4-5): caturndm api varndndm (em. : 

catuvarnndpivarnndnmm Cod.) dcdryatvam ihoditam | brdhmanddicatuskasya 
dvijo 'nugrahakrd bhavet \ ksatriydditrikam yac ca *ksatriyo *diksito (corr. : 
dlksitod Cod.) guruh | vaisyddidvitayam vaisyah sudrah sudrdn tu dlksayet. In this 
[system] the office of Acarya has been taught for all four caste-classes. A brahmin 
may initiate persons of the four beginning with his, an initiated Ksatriya Guru the 
three beginning with his, and a Vaisya the two beginning with his. A Sudra may 
initiate [only] Sudras'. 

Pardkhya quoted in Dlksddarsa A, p. 26; B, p. 42: *amadyapdh (em. : amadyapa A 
: amadyapa B) *kullnds (corr. : kulinas A : kulinan B) ca nityadharma*pardyandh 
(em. : pardyanah AB) | *sudrdh (em. : sudra AB) ksatriyavaj jneyds sesd nindyd<s> 
tato bhrsam 'Those Sudras who do not drink alcohol, who are of good family, and 
devoted to the obligatory religious duties should be looked upon as Ksatriyas. All 
the rest are completely to be condemned'. Cf. Pdramesvara f. 3v2-3: *amadyapds 
(em. : amedhyapds Cod.) tu ye sudrd<h> sau[cd]cdrasamanvitdh | rudrabhaktds 
tu tesdn tu bhojyam annam praklrtitam 'One is permitted to accept food from 
those Sudras who do not drink alcohol, who observe the rules of purity, and are 
devotees of Siva'; Trilocanasiva, Somasambhupaddhativydkhyd, p. 84: tad uktam 
brahmasambhupaddhatau "brahmaksatriyavisdm bhiksdm *abhisastddivarjitdm 
(em. : abhisabdddivarjitdm Cod.) | amadyapds tu ye sudrah saucdcdrasamanvitdh 
| tesdm eva cared bhiksd ndnyesdm tu kaddcana" iti 'That has been taught in the 
Paddhati of Brahmasambhu in the following: "One may gather alms only from brah- 
mins, Ksatriyas, and Vaisyas, provided it is not from someone who been condemned 
[for some sin] or [permanently excluded from his caste], and also from such Sudras 
as do not drink alcohol and observe the rules of purity. One may never accept alms 
from others'". 


Genesis and Development of Tantrism 

of the full set of possibilities, in which members of any caste-class are said to be 
able to initiate only their equals and inferiors, serves, I suggest, not as a record 
that all these possibilities were enacted but rather as an abstraction that adds 
authority to the more restricted common practice by presenting it as following 
a universally valid principle upheld in the brahmanical social system, seen, for 
example in the rule that a man may marry a woman born of parents of his own 
caste or one below it but never a woman from a community ranked above him. 684 
Indeed Saiddhantika texts that discuss who may receive initiation and consecra- 
tion and who may not include the offspring of such forbidden marriages in the 
latter category. 685 

Evidence of the existence of such self-contained Sudra Saiddhantika lin- 
eages is abundant in the Tamil country at the end of our period and after it down 
to modern times. There members of the Sacchudra Vellala community such as 
Meykantar, and Nanacampantar played a significant part in the development of 
the canon of the Tamil Saiva Siddhanta, and a good number of powerful Mathas 
emerged, such as those at Tarumapuram (Dharmapuram) and Tiruvaduturai, 
in which the presiding ascetics were and have continued to be members of this 
upwardly mobile Sacchudra caste. 686 

684 See, e.g., Ydjhavalkyasmrti,Acdrddhydya 57, 91-95. 

685 Diksddarsa A, p. 23; B, p. 25: atrddhikdri*tvanirupanavidhir (corr. : 
nirupanatvavidhir Codd.) ucyate \ viprddindm dasdndm gurutvam uktam | tathd 
cintyavisve "viprddisu catursv evam anulomadisu satsu ca | etesdm dasajdtindm 
dcdryatvam vidhiyate" | tathd kdmike "catvdro brdhmanddyds ca anulomas ca ye 
matdh" 'I shall [now] explain how one determines who is competent for this [office]. 
Ten, beginning with the brahmin, can be Gurus. Thus in the Cintyasiva: "It is ruled 
that these ten castes may be Acaryas: the four beginning with brahmin, and the six 
Anulomas". And in the Kdmika: "The four beginning with the brahmin and the [six] 
Anulomas"'. The term Anuloma here is a synonym of anulomajah 'born of a union 
that is in the natural direction', that is to say, hypergamous. The six Anulomas are 
(1) from a brahmin man and Ksatriya woman (Murdhavasikta), (2) from a brahmin 
man and Vaisya woman (Ambastha), (3) from a brahmin man and Sudra woman 
(Parasava); (4) from a Ksatriya man and a Vaisya woman (Mahisya/Madgu), (5) 
from a Ksatriya man and a Sudra woman (Ugra), and (6) from a Vaisya man and a 
Sudra woman (Karana). See, e.g., Ydjnavalkyasmrti , Acdrddhydya 91-92. The -ddi- 
in anulomadisu satsu ca in the passage cited from the Cintyavisva is redundant and 
may be corrupt (perhaps for anulomdtmasu). 

ese ARE 1909, p. 105; Stein 1994, pp. 235-241; Ghose 1996, pp. 222, 253-282. Stein 
hypothesizes (1994, pp. 237-239) that this rise of the Vellalas was the cause of the 
fact that from the thirteenth century onwards independent shrines of the Goddess 
(kdmakostham) began to be built in the Tamil area alongside those of Siva and to 
be enclosed with the latter in a single architectural complex. He takes this to be 
evidence of "the assimilation of folk conceptions of deity". See also Ghose 1996, 
pp. 221-222. There is certainly widespread evidence of Saktization in the later 
south-Indian Saiva literature. In the south-Indian Saiddhantika scriptures Rau- 
rava, Cintya, Makuta, and Suksma all the male deities in the circuits surrounding 


The Saiva Age 

It might be suspected that this is an isolated development pecular to the Far 
South; and I must say that I am not yet aware of historical evidence of parallel 
developments elsewhere in India at this time or before it. However, it is ex- 
tremely improbable that we would have found unambiguous statements in early 
texts that are very unlikely to have emanated from that region to the effect that 
Sudras may receive consecration as Acaryas, initiate others of their caste and 
pass on their office within it, if this was not indeed a widespread practice. This 
is all the more certain in the light of the fact that the same early corpus provides 
specific instructions on how such initiates should be named, how they should 
dress their hair, mark themselves with ash, and the like. 687 

Siva in temple worship, from the first of the Brahmas to the last of the Weapons 
have been provided with a personal Sakti; see Raurava, Kriydpdda, Patala 59, and 
N.R. Bhatt's introduction to his edition of the Sdrdhatrisatikdlottara, pp. xviii-xix 
(Cintya and Makuta) and pp. lxviii-lxix (Cintya, Makuta, and Suksma). There is 
striking evidence of a related development in the Tamil Saiva literature in the 
Tirumantiram of Tirumular. That text has been assigned to the fifth, sixth, and 
seventh centuries. But it weaves together the Tamil Saiva Siddhanta, the Vedanta, 
a Sakta tradition that features kundaliniyogah and the cult of Tripura, and the cult 
of Nataraja. This is a combination which is unlikely to predate the twelfth century 
(see also Goodall 2004, pp. xxix-xxx). In Sanskrit the same amalgam appears in 
such works as the scripture Jhdnasiddhydgama and the Siddhdntapaddhati of a 
687 Sarvajhdnottara A f. 35r3-5 (14.35-40), B pp. 99-100 (Lihgoddhdrddiprakarana 
vv. 34c-40b): dpddamastakam ydva bhasmasndnam dvijasya tu \ ndbher urdhvam 
nrpasyoktam draktena tu bhasmand || 36 vaisyasya pattikd proktd sudrasya 
tu tripundrakam | bhasmand brahmajaptena yathd*sthdnair (A : sthdnesv B) 
anukramdt || 37 brdhmanasya jatd<h> *suksmd<h> (A : slasnd B) kanakdh 
pariklrtitdh | sthulds taddvigund jneyd ksatriyasya tu vyantardh || 38 vaisyasyaikd 
sikhdsthdne tathd sudrasya kirtitd | hrasvd slaksn*dksasamyuktdh (dksa A : dnu 
B) samyatasya *jitendriya (conj. : jitendriyah Codd.) || 39 *yajnopavitam sautram 
(A : yajhopavltasutram B) tu vipre pahcasaram smrtam | trisaram ksatriyasyoktam 
vaisyasya dvisaram smrtam \\ 40 sudrasyaikasaramjneyam nityam avyabhicdrinah 
| *arcdgnikdryakdle tu (A : arcdydm agnikdrye vd B) samdhydkdle ca ndnyathd ' 
A brahmin's bath with ashes should be from foot to head. A Ksatriya's has been 
taught to be from the navel up and with reddish ash. A Vaisya may have only a 
broad band [of ash] on his forehead. A Sudra may make the Tripundraka marks 
with ash on the various prescribed points on the body in the [prescribed] orderf. In 
each the bath should be done] with ash empowered by the [five] Brahma[mantra]s. 
A brahmin's braids should be narrow and [of the round variety,] called 'thorn apples' 
{kanakdh). A Ksatriya's should be twice as thick *. . . (?). A Vaisya should have only 
one braid, on the crown of his head. It should be short, smooth, with a Rudraksa 
bead attached. The same applies to a Sudra ascetic, *0 you of controlled senses 
(?). The sacred thread should always have five strands for a brahmin, three for 
a Ksatriya, two for a Vaisya, and one for an observant Sudra. The last, however, 
may wear it only when doing Puja, making offerings into the sacrificial fire, and 
during the periods of the junctures of the day'; Kirana f. [60]r3-4 (37.10, 12-13): 
upavitam *bhaved (corr. : bhavedd Cod.) evam ksatriydditrayasya tu | trisaram 
dvisaram vdpi kdryam ekasaram kramdt || 12 pujdtanmdtrakam kdlam nordhvam 


Genesis and Development of Tantrism 

As for those below the Pure Sudras, that is to say, members of Sudra castes 
not considered pure and, below even them, members of the various more or less 
untouchable communities denned as the lowest-born (antyajah), these too were 
drawn by the Saiddhantikas within the reach of the religion. Texts of this tradi- 
tion declare that a Guru is forbidden to give them initiation in the full sacrificial 
form {hautri diksa). But if he sees that they are inspired by sincere devotion to 
Siva he is required to perform for them a simplified form of initiation that avoids 
direct contact. This is to be accomplished mentally {manasi diksa) or in the form 
of a gaze believed to transmit Siva's liberating power {caksusi diksa), or by allow- 
ing them to drink the water with which his feet have been washed, an extension 
of the common devotional practice of drinking the water that gathers at the foot 
of an image in the course of its worship. 688 

tesam bhaved iha \jatdndm dhdranam *bhasmalepanam (corr. : bhasmamlepanana 
Cod.) *brdhmane (corr. : brahmane Cod.) bhavet || 13 tripundraka<m> sikhd caikd 
ksatriydditraye bhavet 'This is how the sacred thread should be [for a brahminl. 
But for Ksatriyas and the rest it should be made with three, two, and one strand 
respectively and may be worn only at the time of worship, not after. A brahmin 
[only] may wear [full] braids and smear [his whole body] with ashes. The three 
[castes] beginning with Ksatriyas may have a Tripundraka and a single [braid 
at the] crown'; Mrgendra, Carydpdda 1.3-4a: vratino jatild mundds tesv agryd 
bhasmapdndardh | tilakaih pundrakaih pattair bhusitd bhumipddayah | jatd na 
sudro bibhrydt 'Ascetics [should either] have their hair in braids or be shaven bald. 
The foremost among them[, the brahmins,] should be white with ash [from head 
to foot]. Ksatriyasf, Vaisyas,] and [Sudras] should be adorned with dots [of ash], 
Vaisyas with the [Tri]pundraka lines, and Sudras with a broad band [of ash on the 
forehead]. A Sudra may not wear braids'. For the differentiation of initiation-names 
according to caste see here p. 291. 
688 Kirana f. [60]v3-4 (38.6c-7): yathdsthitena bhdvena *mantrdh (em. : mantra Cod.) 
kurvanty anugraham || yatas tato *'ntyajasyasydpi (conj. : ntyajasydsydsya Cod.) 
diksa *kim tv atra (em. : kintatra Cod.) manasi | kdrukdndm tu samsparsd<n> 
*na tu hautrim (em. : nugrahautri Cod.) prakalpayet 'Since Mantras grant ini- 
tiation in consideration [only] of the state of [a person's] mentality he may give 
initiation even to an untouchable. But [the initiation] in this case [must be 
only] through the medium of the mind. It the case of workmen [it should be] 
by touching them. He must not do the initiation involving fire-sacrifice [for ei- 
ther]'; Kdmika quoted in the Diksddarsa A, p. 27 and B, p. 43: antyajdndm 
na hautri sydt kim tu diksa tu caksusi 'Untouchables may not receive initiation 
through fire-sacrifice. But they can receive ocular initiation'; Vdyavyasamhitd 
quoted in the Diksddarsa A, p. 26 and B, p. 41: asacchudrdntyajdtindm patitdndm 
visesatah \ tathd samkarajdtindm nddhvasuddhir vidhiyate \ te 'py akrtrimabhdvds 
cec chive paramakdrane | pddodakapraddnddyaih kurydt *pdsavisodhanam (A : 
pdduvisodhanam B) | atrdnulomajdtd ye *yuktd ye (em. : yuktaye AB) *vd (A : va 
B) dvijdtisu | tesam adhvavisuddhyddi *kdryam atra (em. : kdryamdtra AB) *ku- 
locitam (A : kulojitam B) 'The elimination of the paths [of the universe through 
oblations in the sacrificial fire] is not permited for Impure Sudras, untouchables 
(antyajdti-), and, above all, for outcastes (patita-), nor for those of the mixed castes 
(samkarajdti-). Iff, however,] they have genuine devotion to Siva, the highest cause, 


The Saiva Age 

Orthodox brahmanical practice denied all Sudras access through Up- 
anayana to the Veda and the rituals that are animated by its Mantras and 
excluded even more radically the various groups it ranked below these as 'the 
lowest born' (antyajah, antyajatih). The texts of the Saivas justified their liber- 
ating inroads into the mass of humanity beyond these brahmanical boundaries 
by boldly declaring that the system of the separation of the castes (jatibhedah) 
is a fabrication without basis in reality, a cultural epiphenomenon rather than 
a deep fact of nature, 689 pointing to its absence among human beings outside 
of India. 690 Only mentality matters; and consequently all devotees of Siva 
form a single community regardless of birth, 691 one whose only true internal 

he should eliminate their bonds by such means as giving them the water from his 
feet. As for those who are born of inter-caste marriages in which the father's caste 
is higher or *if they are connected with brahmins (?) he may do [for them the full 
ritual procedure] that begins with the elimination of the paths as appropriate to 
the [caste of the] family [in which they have been born]'. The term samkarajdtih, 
which I have translated literally as 'of the mixed castes' refers to offspring of such 
unions as that between a Mahisya (born of a Ksatriya man and Vaisya woman) and 
Karana woman (born of a Vaisya man and Sudra woman; see, e.g., Mitdksard on 
Ydjhavalkyasmrti, Acdrddhydya 95. 

689 Pauskarapdramesvara quoted in Nityddisamgraha f. 62vl2-13: manusyajdtir 
ekaiva 'There is only one caste, that of human beings'; f. 63r4-5: na jdtir vihitd 
tatra varnam vdpi sitddikam | yonilihgodbhavdh sarve jiva ekah samah sthitah | 
tatra sarvagato devo drsyate jndnacaksusd | ajhdna*dhvastacittdndm (conj. : pdpa- 
cittdndm Cod.) kusdstra*vivasdtmandm (conj. : vihitdtmandm Cod.) | vdkpraldpah 
sthitas tesdm yadi jdtih prayojanam 'No caste has been enjoined with respect to 
them, nor colour such as white. All are born from sexual union and the souls [of 
all] are equal. With the eye of knowledge Siva is seen pervading all of them. If 
[they declare that] caste is relevant then this is the prattling of men whose un- 
derstanding is destroyed by ignorance, who are under the sway of false teachings'; 
Kulasdra f. 72r2: ekabljaprasutam hi sarvam jagad idam priye | tasmdj jdtivicdram 
tu bhrdntipurvam idam krtam 'This whole world, my beloved, has been born from 
a single seed. So this concern for caste that people have springs from an error'; 
Tantrdloka 15.595c-601b. 

690 Cintyavisvafsdddkhya] quoted in Diksddarsa of Vedajnanaguru II, A, p. 24; B, p. 38: 
navakhandesu sarvesu bhdratesu *mayena ca (B : ca yena ca A) \jdtibhedam idam 
kalpyam anyadesesu ndsti tat \ tasmdt tat kalpandmdtram jdtibhedam *iti kramam 
(?) 'Maya [the Guru of the Asuras] created this division of the castes throughout 
the nine divisions of the continent of Bharata. It does not exist in other countries. 
Therefore it is nothing but a fabrication/fiction.' 

691 See, for example, Nisvdsakdrikd, pp. 35-36 (12.161-167): 161 tattvdni yo vijdndti 
tattvdndm *vydptim uttamdm (em. : vydptir uttamam Cod.) | dharmddharmdn 
na lipyeta sa sarvdnugrahe ksamah || 162 brdhmana<h> ksatriyo *vaisyah (corr. 
: vesyah Cod.) sudro vd tattvavid yadd | *vibhaktir (em. : vibhaktim Cod.) 
naiva vidyeta yathdgndv agnir eva hi || 163 ksiram ksire yathd nyasto toye toyam 
ivdrpitam | vibhdgo naiva vidyeta tattvam lsvara*bhdsitam (conj. : Tsvarabhdvitam 
Cod.) || 164 yathd hi saritas sarvds sdgardsrayasamsthitdh | *vivektum (em. : 
vivekan Cod.) tu na sakyante rasa*bhdve (conj. : bhdvam Cod.) prthak prthak 
|| 165 tadvad varndsramd devi diksito yadi vd pasuh | sivabhdvasamd*yuktds 


Genesis and Development of Tantrism 

hierarchy is that created by the four levels of empowerment through initiation 
and consecration. 692 

However, it should not be imagined that because they insisted that the divi- 
sions of the castes are ultimately groundless when explaining the inclusiveness 
of their recruitment they rejected these divisions in practice. It is one thing to 
extend one's recuitment into lower social strata and quite another to reject the di- 
visions between them in practice. Thus in spite of their rhetoric of the underlying 
unity of man they required that caste divisions be respected not only in relations 
between initiates and the wider society in matters such as marriage but also in 
relations between fellow-initiates. As we have seen, they denied impure Sudras 
and untouchables the full ceremonial form of initiation, they refused to transmit 
the office of Acarya to the offspring of unions between a man of a lower caste 
and a women of a higher, and they would not countenance an Acarya's initiating 
his caste superior, in effect a Sudra's initiating a brahmin. They also required, 
for example, that when initiates of different caste-classes gathered they should 
sit apart, each in a separate line; 693 the penances (prayascittam) that they pre- 

(conj. : yukto Cod.) *tulya (conj. : tulyam Cod.) eva na samsayah || 166 
sivatantram samasritya vibhaktim yah karisyati | *pacyen narah sa (conj. : sa 
pacyen naro Cod.) ghoresu dvatrimsan narakesu ca || 167 brahmanas tu dinah 
pahca dinah pahca ca kesave | dinatrayam tu rudrasya pray ascittly ate narah; 
Valadharin, Kriyasamgrahapaddhati f. 49r4-vl, extending this principle to in- 
clude foreigners (better to initiate a sincere Mleccha than an insincere brah- 
min): mayanvito yada sisyo viprajatisamudbhavah | mayahinas tatah patram 
mlecchasudradisambhavah || na vipre dapayed dlksam dapayen mlecchajanmine 
| nadhikari yato vipro mayadigunasamyutah \\ nisprapancagunair yukto mlecchas 
caiva sivagame | dlksa vai sarvatha tasya yato mayavivarjitah. See the same point 
made in the lost scripture Mukuta cited by Jayaratha on Tantraloka 15.514cd. 

692 Nityadisamgraha f. 63rll-12: taponibaddho yair atma brahmanams tan vidur 
janah | pasupasavidhanajnah sivajnananusarinah | te hi devatidevasya puja- 
karmani kirtitah | ity uktam candrahasakhye mukutadyagamesu ca samayya- 
divisesena jatir ekaiva kirtita 'People judge as [true] brahmins those who have con- 
trolled themselves through austerity, who know the bound soul, the bonds, and the 
rites [of initiation], and who follow the teachings of Siva. For it is these that have 
been declared [fit to officiate] in the rites of the worship of the Supreme Deity. This 
has been taught in the [scripture] Candrahasa'; and in such texts as the Mukuta 
we are told that there is only one 'caste' [for Saivas] with differentiation [by status] 
only into Samayinsf, Putrakas, Sadhakas,] and [Acaryas]'. 

693 Somasambhu, Brunner 1961, p. 301 (v. 8cd.): savarnair ekaya pahktya 
bhunjitantarmanamunih 'One should eat in silence with concentrate mind in a sin- 
gle line with others of the same caste-class'; Trilocanasiva, Prayascittasamuccaya, 
p. 25: ekapahktih sada varjya bhojane bhinnajatibhih 'When eating one must al- 
ways avoid sitting in a single line with persons of other castes'. Note the distaste 
expressed by the brahmin Samkarsana in the Agamadambara (p. 56) when, in a 
Kashmirian monastery, he notices that Buddhist monks do not form separate lines 
according to caste when they eat together: catvaro varna varnasamkara api va 
sarva evaikasyam pahktau bhunjate 'Persons of all the four caste-classes and even 


The Saiva Age 

scribed for initiates contaminated by an accidental or wilful contact with a per- 
son in a state of impurity were calibrated in severity according the degree of dis- 
tance in caste-status between the persons contaminating and contaminated; 694 
and they assigned compound initiation-names such as Aghora-siva and Aghora- 
gana whose second member indicated the caste-status of the bearer, marking 
out brahmins from non-brahmins, non-Sudras from Sudras, or each of the four 
caste-classes from each other. 695 

from the mixed castes are eating together in a single line'. 

See Trilocanasiva, Prayascittasamuccaya p. 25. Similar differentiation accord- 
ing to caste applies to the penances for eating the leavings of another's food 
(ucchistabhojanam), illicit sexual intercourse, and the taking of human life; see 
ibid., pp. 32, 35, 48, and 52-53. How the hierarchy of caste was perceived in rela- 
tion to that between the initiated and the uninitiated can be seen in the rules for 
the penances needed to restore purity if one's food has been contaminated through 
contact with an ucchistah, a person who has eaten but has not yet purified himself. 
The rules for initiated brahmins will suffice to illustrate this. If a brahmin initiate's 
food is contaminated by another brahmin initiate the penance is 100 repetitions of 
the Tatpurusa, the Mantra that is the Lord of his Caste (jdtisah). It is doubled if 
the contaminator is an uninitiated brahmin or an initiated Ksatriya. One day of 
fasting is added to the repetitions if the contaminator is an uninitiated Ksatriya, 
two if the contaminator is an initiated Vaisya, three if an initiated Sudra, four if an 
uninitiated Vaisya, and six if an uninitiated Sudra (ibid., p. 31). Here we see traces 
of a view that the status bestowed by Saiva initiation should prevail over that of 
caste. In its pure form this would entail that a Saiva brahmin should consider con- 
tamination by an initiated Sudra one degree less severe than that by an uninitiated 
brahmin, two degrees less severe than that by an uninitiated Ksatriya, and so on. 
But the Saiddhantikas have preferred to limit the application of this view to the 
lowest two castes, where it was of least consequence, allowing an initiated Sudra to 
be less contaminating than an uninitiated Vaisya, but not a initiated Vaisya to out- 
rank an uninitiated Ksatriya or an initiated Ksatriya an ordinary brahmin. In other 
words the primary distinctions here are (1) that between brahmins and Ksatriyas 
on the one hand and Vaisyas and Sudras on the other, and (2) that between brah- 
mins and Ksatriyas. So while a Sudra will be purer than a Vaisya if he has been 
initiated, a Ksatriya, in effect the king or a member of his family, will never be less 
pure than a Vaisya, nor a brahmin less pure than a non-brahmin. In this regard 
the benefit of initiation in the case of the Ksatriya is limited to an acceptance that 
he is no more contaminating than an uninitiated brahmin. But this is already a 
major concession in terms of caste and articulates the view seen elsewhere in the 
literature that the prosperity of society requires an alliance between the brahmins 
led by the Saivas and a monarch who has received initiation from the Saiva Guru. 
This view is underlined by the fact that penance is without fasting in the case of 
contamination by brahmins or an initiated Ksatriya but with fasting in all other 

I am aware of five different rulings in this matter. (1) names in -siva, etc. for brah- 
mins only, in -gana for Ksatriyas, Vaisyas, and Sudras, and in -sakti for women; 
see Kirana 37.11-12b: ksatriydditrayasyoktam <m>antranama ganahkitam || 12 
vipranam *mantrapurvam (conj. : matupurvvan Cod) tu sagotrantam bhaved iha 
'In the case of the three [castes] beginning with the Ksatriya it should be the name of 
one of the Mantras distinguished by [the addition of] -gana. In the case of brahmins 


Genesis and Development of Tantrism 

However, the non-Saiddhantika traditions of the worship of Bhairava and 
the Goddess in the Mantrapitha and Vidyapitha have shown themselves much 
less willing to tolerate such compromises, seeing them as a contamination of the 
true Saiva tradition and appropriate only for those, namely the Saiddhantikas, 
whose degree of illumination by Siva is insuffficient to enable them to appreciate 
and enact his higher teachings. 696 Distinction on the basis of caste is generally 

it should begin with a Mantra and end with the Gotra name [-siva, etc.]'; Mrgendra, 
Kriydpdda 8.60c— 61: srajam vimocayen ndma diksitdndm tadddikam || sivdntakam 
dvijendrdndm itaresam gandntakam 'He should throw the garland. The names of 
initiated brahmins should begin with [the name of] that [on which it lands] and end 
in -siva. For all others it should end in -gana; 'and Vidydpurdna, a Saiddhantika 
scripture in spite of its title, quoted in Nityddisamgraha f. 63vl2- 64rl3: sivo 
jyotih sikhd caiva sdvitras ceti gocardh \ ... etdh samjhd dvijdgrydndm rdjddlndm 
gandhkitdh \ saktisamjhds tu *vai (em: vd) strindm sarvdsdm parikirtitdh 'The 
gocaras are Siva, Jyoti, Sikha and Savitra. . . . These names [ending in -siva, 
-jyotis etc.] are proper to brahmins. The names of Ksatriyas[, Vaisyas] and 
[Sudras] are distinguished by the [ending] -gana, while all women are required 
to have names [ending] in -sakti'; (2) a Kashmirian tradition in which names 
in -siva are for the three higher caste-classes, with names in -gana for Sudras 
only, and names in -sakti for women; see Bhatta Narayanakantha on Mrgendra, 
Kriydpdda 8.60c-61 cited above, taking dvijendrdndm there to mean not brahmins 
but brahmins, Ksatriyas, and Vaisyas; Jayaratha, Tantrdlokaviveka on 4.265ab 
(adding names in -sakti for women); Manoda, Kalddlksdpaddhati A ff. 96vl6- 
97r9: tatpdtdvasare sivandmdhkitam sisyam vidhdya striyam ca saktindmdhkitdm 
vidhdya . . . sudravisaye tu ayam amukagana dgatah iti prayojyam 'When that 
[flower] falls he should name a male disciple -siva and a woman -sakti. ... In the 
case of a Sudra he should formulate [the Mantra] as follows: 'This man, N-gana, has 
come [before you, O Lord]'; (3) names in -siva for brahmins, and in -gana and -deva 
for Ksatriyas and Vaisyas; see Brahmasambhu, Naimittikakarmdnusamdhdna 
f. 38v4-5 (2.180): tatpdtasucitasthdnapurvam sivapadottaram | ndmdvadhdrya 
viprasya ganadevdntam anyayoh 'Having determined the [initiation] name, whose 
first part should be the * ...(?) indicated by the fall of the [flower] and whose 
second part should be the word -siva in the case of a brahmin, but which should 
end in -gana and -deva in the case of the other two [castes]'; Amrtesadiksdvidhi 
f. 16r6-7: sisyasya namakaranam sivdmaragandntakam; (4) names in -siva for 
brahmins, and in -deva, -gana, and -muni for Ksatriyas, Vaisyas, and Sudras; see 
Isdnasivagurudevapaddhati, Kriydpdda 146 (16.67-68b): sivdntam brdhmanasya 
sydd devagandntam anyayoh | sudrasya munisabddntam ndma kurydd 'The name 
of a brahmin should end in -siva and those of the next two [castes, Ksatriya and 
Vaisya] in -gana and -deva. He should give a Sudra a name that ends in -muni'; 
and (5) names in -siva for brahmins, -kavaca for Ksatriyas, -deva for Vaisyas, 
and -gana for Sudras; see Brhatkdlottara A, f. 91v3-4: sivasamjnd dvijasyaiva 
kavacakhya nrpasya ca | vaisydndm devasamjnd ca sudrdndm ca *gandntakam (em. 
: gandntikam Cod.) | puspapdtdnusdrena